Cloud Object Storage

User Guide

Product Documentation

Copyright Notice

©2013-2025 Tencent Cloud. All rights reserved.

Copyright in this document is exclusively owned by Tencent Cloud. You must not reproduce, modify, copy or distribute in any way, in whole or in part, the contents of this document without Tencent Cloud's the prior written consent.

Trademark Notice

All trademarks associated with Tencent Cloud and its services are owned by the Tencent corporate group, including its parent, subsidiaries and affiliated companies, as the case may be. Trademarks of third parties referred to in this document are owned by their respective proprietors.

Service Statement

This document is intended to provide users with general information about Tencent Cloud's products and services only and does not form part of Tencent Cloud's terms and conditions. Tencent Cloud's products or services are subject to change. Specific products and services and the standards applicable to them are exclusively provided for in Tencent Cloud's applicable terms and conditions.

Contents

User Guide

Creating Request

Overview

Last updated：2025-11-18 16:05:05

Concept

Tencent Cloud COS is a web-based storage service accessed using the HTTP/HTTPS protocol. You can use RESTful APIs or COS SDKs to access COS.
Your COS access request must first pass the COS identification and authentication before COS starts to operate the resources. Therefore, depending on whether the identity is identifiable, COS access requests are divided into two types: anonymous request and signed request.
Anonymous request: If a request does not include Authorization or related parameters, or the user identity cannot be identified based on the related characteristics, the request will be treated as an anonymous request for authentication.
Signed request: A signed request must contain the Authorization field in the HTTP header or the request packet. The content of the field is generated based on Tencent Cloud security credentials (SecretID and SecretKey) and some characteristic values of the request via an encryption algorithm.
To access COS using COS SDKs, you only need to configure your security credentials before initiating the request. To access COS using the REST API, calculate the request signature according to Request Signature.

Obtaining Security Credentials

Cloud Access Management (CAM) provides features and services related to accounts and credentials for COS, to help customers manage the permissions to access resources under their Tencent Cloud accounts in a secure way. You can use CAM to create, manage and terminate users (or user groups), and manage other users' permissions to use Tencent Cloud resources through identity management and policy management.

Security credentials of the root account

After logging in to the root account, you can manage and obtain the security credentials (SecretID and SecretKey) of your root account on the Cloud API Key page of CAM. The following is a key pair example:
36-character access key ID (SecretID): ************************************
32-character access Key (SecretKey): ********************************
The access key can be used to identify the uniqueness of an account. After the signature is generated using the key and the request is sent, Tencent Cloud will identify the identity of the request initiator, and then perform verification and authentication for the identity, resources, operations, and conditions to determine whether to allow the operation.
Note:
The key of the root account has all the operation permissions for all resources under the root account. Disclosure of the key may cause loss of your cloud assets, so it is strongly recommended that you create sub-accounts and assign corresponding permissions for them, and then use the keys of sub-accounts to create requests for resource access and management.

Security credentials of sub-accounts

To manage users and cloud resources under your account in multiple dimensions, you can create multiple sub-accounts under your primary account to implement user-specific permission management. For more information on how to create a sub-account, see Sub-users in CAM.
Before using a sub-account to initiate an API request, you need to create a security credential for the sub-account, and then the sub-account will get a unique key pair, which can facilitate the identification of the identity. You can create user policies for different sub-accounts to control their access permissions to resources. You can also create user groups and associate one access policy to a user group to facilitate the central management of user grouping and resources.
Note:
With the corresponding permissions assigned, a sub-account can create or modify resources. The resources still belong to the primary account, and the resource cost will be deducted from the root account.

Temporary security credentials

In addition to using security credentials of the root account or sub-accounts to access resources, you can create roles and use the temporary security credentials of the roles to manage your Tencent Cloud resources. For more information on the role concept and how to use roles, see Role Overview.
As a virtual identity, a role does not have a permanent key. Tencent Cloud CAM provides a set of STS APIs used to generate temporary security credentials. For more information on how to use the APIs and relevant examples, see Using Roles. See CreateRole to learn about how to generate temporary security credentials. Temporary security credentials contain only limited policies (operations, resources, and conditions), and are valid for a limited period (start and end time), so the generated temporary security credentials can be distributed or used directly.
You can call the API for generating temporary security credentials and get a temporary key pair (tmpSecretId/tmpSecretKey) and a security token (sessionToken), which form the security credential that can be used to access COS. The following is an example of a temporary security credential:
41-character security token (SecurityToken): *****************************************
36-character temporary access key ID (SecretID): ************************************
32-character temporary access key (SecretKey):********************************
This API also returns the validity period of the temporary security credential via the expiration field, which means that this set of security credentials can only be used to initiate requests during this period.
Tencent Cloud COS provides a simple server-side SDK that can be used to generate temporary keys. You can visit COS STS SDK to obtain the SDK. To initiate the request using the REST API after getting the temporary security credential, you need to specify the value for the x-cos-security-token field in the HTTP header or the form-data of the POST request packet to identify the security token used by the request, and then use the temporary access key pair to generate the request signature. For more information on how to initiate requests using the COS SDK, see the relevant sections in each SDK documentation.

Access Endpoint Domain Names

RESTful APIs

The Region and Access Endpoints document provides a list of domain names that can be used to initiate access requests via RESTful APIs.
We recommend you use virtual hosting domain names to access COS buckets. When you initiate an HTTP request, the bucket to be accessed will be specified through the Host header, for example, <BucketName-APPID>.cos.<Region>.myqcloud.com. Using virtual hosting domain names implements the same feature as the root directory of a virtual server. Virtual hosting domain names can be used to host files such as favicon.ico, robots.txt, and crossdomain.xml, which are the content that many applications will retrieve from the root directory of the virtual server by default when identifying a hosted website.
You can also use a path request to access a bucket, for example, cos.<region>.myqcloud.com/<BucketName-APPID>/. The request Host and the signature must use cos.<region>.myqcloud.com. COS SDKs do not support this access method by default.

Domain names of static websites

If you enable the static website feature for a bucket, a virtual hosting domain name will be assigned for you to use relevant features. Unlike RESTful APIs, the domain name of a static website supports only a few operations, such as GET/HEAD/OPTIONS Object, in addition to specific index pages, error pages and redirection configurations. Uploading or configuring resources is not supported.
The format of a domain name of a static website is <BucketName-APPID>.cos-website.<Region>.myqcloud.com. You can also log in to the console and go to the bucket's Basic Configuration > Static Website Configuration to get the domain name.


Private network access

Tencent Cloud COS adopts intelligent resolution for COS endpoints. In this way, the optimal linkage can be provided for you to access COS with different ISPs.
If you have deployed a CVM within Tencent Cloud for accessing COS over a private network, you must first ensure that the CVM resides in the same region as the COS bucket, then use the nslookup command on the CVM to resolve the COS endpoint. If a private IP is returned, access between the CVM and COS is over a private network; otherwise, it is over a public network.
If your CVM resides in a different region from the COS bucket, but it is still in one of COS regions, you can use the COS private network global acceleration domain name to access files and achieve cross-region access between the CVM and COS.

How to determine whether access is via a private network

Tencent Cloud products within the same region can access each other over a private network, incurring no traffic fees. Therefore, we recommend choosing the same region when you purchase different Tencent Cloud products to save on costs.
Note:
The private networks of Public Cloud regions do not interconnect with those of Finance Cloud regions.
The following shows how to determine access over a private network:
For example, when a CVM accesses COS, to determine whether a private network is used for access, use the nslookup command on the CVM to resolve the COS endpoint. If a private IP is returned, access between the CVM and COS is over a private network; otherwise, it is over a public network.
Note:
Generally, a private IP takes the form of 10.*.*.* or 100.*.*.*, and a VPC IP takes the form of 169.254.*.*. These two types of IPs belong to private networks.
Assume that examplebucket-1250000000.cos.ap-guangzhou.myqcloud.com is the address of the destination bucket. After running the nslookup command, you can view the information as shown in the figure below.



In the command output, the 10.148.214.13 and 10.148.214.14 IPs indicate that the access to COS is over a private network.

Testing connectivity

Basic connectivity test

COS uses the HTTP protocol to provide services. You can use the most basic tool telnet to test the connectivity to port 80 of the COS access domain.
The following is an example of access through the public network:
telnet examplebucket-1250000000.cos.ap-guangzhou.myqcloud.com 80
Trying 14.119.113.22...
Connected to gz.file.myqcloud.com.
Escape character is '^]'.
The following is an example of access through Tencent Cloud CVMs (classic network) within the same region:
telnet examplebucket-1250000000.cos.ap-guangzhou.myqcloud.com 80
Trying 10.148.214.14...
Connected to 10.148.214.14.
Escape character is '^]'.
The following is an example of access through Tencent Cloud CVMs (VPC) within the same region:
telnet examplebucket-1250000000.cos.ap-guangzhou.myqcloud.com 80
Trying 169.254.0.47....
Connected to 169.254.0.47.
Escape character is '^]'.
Regardless of the access environment, if the command returns the Escape character is '^]'. field, it indicates that the connection is successful.

Test via the internet

The access to COS over the internet involves the ISP network, which may prohibit you from testing connectivity using tools such as ICMP ping or traceroute. Therefore, we recommend you use TCP tools to test connectivity.
Note:
The access via the Internet may involve multiple network environments. If the access is not smooth, check your local network linkage, or contact the local ISP.
If your ISP allows you to use the ICMP protocol, you can use the ping, traceroute or mtr tools to check your linkage. Otherwise, you can use the psping (Windows environment; download at the Microsoft official website) or such tools as tcping (cross-platform software) to test the latency.

Test via a private network

If you access the COS over the Tencent Cloud VPC in the same region, you may be unable to test connectivity using tools such as ICMP ping or traceroute. We recommend that you use the telnet command in the basic connectivity test to perform the testing.
You can also use tools such as psping or tcping to test the latency to port 80 of the access domain. Before the test, make sure that the access domain name has been correctly resolved to the private network address using the nslookup command.

Bucket

Bucket Overview

Last updated：2025-10-11 17:41:47

Overview

A bucket is the carrier of objects, which can be understood as the "container" for storing objects, and this "container" has no upper limit of capacity. Objects are stored in buckets in a flat structure with no concept of folders and directories. You can choose to store objects in one or multiple buckets.
Note:
A bucket can contain any number of objects, but one root account can create only up to 200 buckets.

Bucket Naming Conventions

A bucket name consists of BucketName and APPID connected by a hyphen (-). The naming format for the bucket is <BucketName-APPID>.
BucketName: A string of characters manually entered by the user, with the following naming conventions:
Only lowercase letters (a-z), digits (0-9), and hyphens (-) are allowed.
The number of characters a bucket name is allowed to contain is limited by the length of the region abbreviation and APPID. The combined domain name can be up to 60 characters. For example, the domain name 123456789012345678901-1250000000.cos.ap-beijing.myqcloud.com contains 60 characters.
A bucket name cannot start or end with "-".
APPID: This is the account obtained upon successful application for a Tencent Cloud account, automatically assigned by the system with fixed and global uniqueness. It can be viewed in Account Info. When creating a bucket via console, no user input is required, but specifying the APPID is necessary when using tools, APIs, or SDKs.

Naming Example

For example, the bucket examplebucket-1250000000, where examplebucket is the BucketName (a string manually entered by the user) and 1250000000 is the APPID (automatically assigned by the system when users apply for a Tencent Cloud account).
The following are examples of valid bucket names:
examplebucket-1-1250000000
mybucket123-1250000000
1-newproject-1250000000

Region to Which the Bucket Belongs

Region is where the COS IDC is located. COS allows users to create buckets in different regions. You can select the region closest to the location where you deploy your business for the buckets so as to reduce latency and cost, and meet the compliance requirements.
For example, if your business is distributed in South China, creating buckets in the Guangzhou region can accelerate object uploads and downloads. For more information on regions, see Regions and Access Endpoints.
Note:
Region must be specified when a bucket is created and cannot be modified once specified. All objects in the bucket are stored in the IDC in the region. You cannot set regions for objects.

Bucket Permission Types

A bucket provides two types of permissions by default: public and user.
Note:
If the bucket permission is private read/write or a specified account is granted the permission, an object request needs to carry a signature for identity verification. For more information on signature, see Request Signature.
If the bucket permission is public read/private write or public read/write, an object request doesn't need to carry a signature, and anonymous users can directly access the object at the URL. However, your data may be leaked. Therefore, proceed with caution.

Public permissions

Public permissions include "Private Read/Write", "Public Read/Private Write", and "Public Read/Write". You can modify bucket access permissions in Permission Management of the bucket in the COS console. For more information, see Basic Concepts of Access Control.
Private Read/Write Only the creator of the bucket and authorized accounts have Read/Write permission on the objects in the bucket. The default access permission of a bucket is Private Read/Write, which is recommended.
Public Read/Private Write Anyone (including anonymous visitors) has Read permission on the objects in the bucket, but only the bucket creator and authorized accounts have Write permission on them.
Public Read/Write Anyone (including anonymous visitors) has Read/Write permission on the objects in the bucket, which is not recommended.

User permissions

A root account has all the permissions (full access) for buckets by default. In addition, you can add sub-accounts that are granted permissions to read/write data and permissions, and even full access.

Usage

Viewing Bucket Overview Via Console

Note:
If the sub-account does not have permission to access the dashboard, contact the root account and ask for request permission by adding user policy GetBucket.
1. Log in to the COS console and click Bucket List on the left sidebar.
2. Locate the bucket you want, and click the bucket name.
3. Click the Overview tab to enter the bucket overview page.

Usage Overview

Usage Overview shows the number of objects, incomplete multipart uploads, storage usage, traffic, and requests in the current bucket.
Note:
"Usage Overview" data is delayed for about two hours compared with real-time data. It is for monitoring purposes only. For accurate billing data, go to Billing Center to download usage details.
Number of objects/Number of incomplete multipart uploads: allows you to view the number of objects or incomplete multipart uploads in each COS storage class.
Storage amount: View storage usage for various storage types. Currently supported storage types include STANDARD, MAZ_STANDARD, STANDARD_IA, MAZ_STANDARD_IA, ARCHIVE, MAZ_ARCHIVE, DEEP_ARCHIVE, INTELLIGENT_TIERING, and MAZ_INTELLIGENT TIERING.
Note:
MAZ_STANDARD, MAZ_STANDARD_IA, MAZ_ARCHIVE, DEEP_ARCHIVE, INTELLIGENT_TIERING, and MAZ_INTELLIGENT TIERING are only supported in specific regions. For supported regions, see Storage Type Introduction.
Traffic: View month-to-date total traffic, public network downstream traffic, private network downstream traffic, CDN origin-pull traffic, and global acceleration downstream traffic for various storage types. Currently supported storage types include STANDARD, MAZ_STANDARD, STANDARD_IA, MAZ_STANDARD_IA, INTELLIGENT TIERING, and MAZ_INTELLIGENT TIERING.
Requests: Allow you to view the number of all requests, read requests and write requests for the current month.
Retrievals: Allow you to view the amount of data retrieved from STANDARD_IA and ARCHIVE.

Basic Information

Basic Information includes bucket name, region, creation time and access permissions.
Bucket Name: consists of a custom bucket name and APPID. For naming information, see Naming Conventions.
Region: specifies the region where the bucket resides.
Creation Time: specifies the time when the bucket was created.
Access Permissions: Specify the access permissions for the bucket. For more information on permissions, see Setting Access Permission.

Domain Information

Domain Information shows all the domain names/endpoints configured for the bucket.
Endpoint: the default COS access endpoint for this bucket. It is auto-generated based on the bucket’s name and region when you create a bucket.
Note:
The default domain name may pose security risks and possibly expose sensitive information. It is recommended to configure a custom domain name for your bucket. For details, see Bucket Custom Domain Name Switchover.
Default CDN acceleration domain: The domain name via CDN acceleration nodes is generated by default and can be configured manually by the user. If this feature was previously enabled, this option will be displayed. For more information, see Default CDN Acceleration Domain Name.
Custom CDN acceleration domain: Allows you to bind a custom domain name to Tencent Cloud CDN to speed up access to the objects in this bucket.
Custom endpoint: Allows you to bind your own domain name as a custom endpoint to the bucket for access to the objects in it.
Global acceleration endpoint: The auto-generated endpoint after you enable global acceleration. You can use this endpoint to speed up uploads to the bucket globally. For more information on global acceleration, please see Global Acceleration > Overview.
Static website endpoint: Allows you to access a bucket configured as a static website. For more information on static website, please see Setting Up a Static Website.

Bucket Configuration

Bucket Configuration shows the status of each bucket configuration.
MAZ Configuration: The multi-AZ storage architecture offered by COS, which can provide IDC-level disaster recovery capabilities for your data. For more information, see MAZ Feature Overview.
Metadata Acceleration: A high-performance file system feature provided by COS. For more information, see Metadata Acceleration Overview.
CORS: Represents cross-origin access which requests resources from another origin over HTTP. Two origins that differ in any one of protocol, domain name, and port are treated as different origins. For more information, see Setting Cross-Origin Access.
Versioning: Retains multiple versions of an object after you enable versioning on your bucket. It helps to retrieve your data lost due to accidental deletion or application failure. For more information, see Setting Versioning.
Origin-Pull: Leads you from COS to another origin for data access using an origin-pull rule when the object you request does not exist in the COS bucket or a specific request needs to be redirected. For more information, please see Setting Origin-Pull.
Bucket encryption: By setting bucket encryption, you can encrypt all new objects uploaded to a bucket with the specified encryption method by default. For more information, see Bucket Encryption Overview.
Inventory: The inventory feature can generate inventory reports about object attributes, configuration details, and more for your bucket on a daily, weekly, or monthly basis. For details, see Setting Inventory.
Hotlink Protection: Prevents malicious programs' cheating for public network traffic using resource URLs or stealing of resources by malicious means. For details, please see Setting Hotlink Protection.
Lifecycle: Automatically transitions or deletes specified objects within the specified time according to your lifecycle rule. For more information, see Setting Lifecycle.
Cross-Bucket Replication: Automatically replicates incremental objects asynchronously from the source bucket to the destination bucket in another region after you enable cross-region replication. For more information, see Setting Cross-Bucket Replication.
Logging: By using the logging feature, COS can record various request logs related to bucket operations to help you better manage and use your bucket. For details, see Log Storage Settings.
Tag: a bucket tag used as an identifier to help group and manage buckets. For more information, see Setting Bucket Tags.

Alarm Configuration

Alarm Configuration enables you to configure alarms on your bucket for daily monitoring purposes.
Current alarms: Show the number of ongoing alarms.
Alarm policies: Show the number of existing alarm policies.

Other Configurations

In addition to COS configuration, COS also integrates Cloud Infinite (CI) features, such as content review and data processing. For more information, see Content Review and Data Processing Overview.

Notes

COS stores objects using a flat structure instead of folders. For more information, see “Folders and Directories” in Object Overview.
Each root account (i.e., the same APPID) can create up to 200 buckets in total in all regions. There is no limit on the number of objects in a bucket.
In Tencent Cloud, the bucket name under one APPID is globally unique and must be unique.
Once a bucket is created, it cannot be renamed. To rename a bucket, you need to delete it and create another one with the desired name.
When creating a bucket, make sure to select the desired region, as the region cannot be changed once specified.

Creating a Bucket

Last updated：2025-10-13 10:31:27

To store files in Cloud Object Storage (COS), you need to create a bucket to store objects. You can create a bucket via the console, tools, API, or SDK. For the bucket concept, please refer to Bucket Overview. The following is a detailed introduction on how to create a bucket.
After creating a bucket, you can upload objects to it and configure other features for it, such as setting up a static website, setting bucket tags, and setting bucket encryption. For more configuration instructions, see Console Overview.

Description

One root account can create up to 200 buckets.
Once a bucket is created successfully, its name and region cannot be modified.
Bucket names must be unique under the same root account and cannot be changed.
Bucket names can only contain lowercase letters, numbers, and hyphens, i.e., [a-z, 0-9], and their combinations. The maximum allowed characters for a bucket name are influenced by the region abbreviation and APPID, with a total of up to 60 characters for the complete request domain name. For example, the request domain 123456789012345678901-1250000000.cos.ap-beijing.myqcloud.com has a total of 60 characters.
A bucket name cannot start or end with "-".

How to Use

Using COS console

1. Log in to the COS Console.
2. In the left sidebar, click bucket list.
3. On the bucket list page, click Create Bucket.
4. In the pop-up Create Bucket dialog box, configure the following information:
4.1 Information
Region: Please select the COS region corresponding to the physical area where your business (or number of users) is concentrated. This setting cannot be modified after creation. For more information, see regions and access endpoints.
Name: Enter a custom bucket name. It cannot be modified after creation. For naming specifications, see naming specification for buckets.
Access Permission: Buckets provide three types of access permissions by default: Private Read/Write, Public Read/Private Write, and Public Read/Write. These can still be modified after setting. For details, see bucket access permissions.
Endpoint: Automatically generated. After bucket creation, you can use that domain name to access the bucket.
MAZ configuration: The Multi-AZ feature is an identifier for buckets. When you enable the multi-AZ configuration, your data will be stored in different IDCs within the same region, providing intra-city disaster recovery. Currently, this feature is applicable only to partial regions such as Beijing, Shanghai, Guangzhou, Hong Kong (China), and Singapore region. For more applicable regions and feature introduction, see the Multi-AZ feature overview document.
Note:
Multi-AZ storage incurs higher storage capacity fees than single-AZ. For details, see product pricing in each region.
After enabling multi-AZ configuration for a bucket, it cannot be modified. Once enabled, data will be stored in the bucket with storage types that support multi-AZ features (such as MAZ_STANDARD and MAZ_STANDARD_IA). Configure with caution. If version control is also enabled, its status cannot be modified.
The COS bucket you created cannot enable multi-AZ configuration. It can only be enabled when creating a new bucket.
4.2 Advanced optional configuration
Note:
This setting item is selectable, configure as needed.
Versioning: Once enabled, uploading objects with the same name or performing operations like add, delete, modify will save earlier versions, convenient for you to retrieve object history.
Note:
If version control and the Multi-AZ feature are enabled simultaneously, the version control state cannot be changed. Configure with caution.
Metadata Acceleration: This configuration item is currently only applicable to allowlisted users. For details, see overview of metadata acceleration function.
Smart Compression: Once enabled, each time you access images of specified format in the bucket, they will be compressed in real time without changing the image format, drastically saving traffic. For details, see introduction to smart compression.
Logging: Records various request logs related to bucket operations for you.
Bucket Tag: A bucket tag serves as an identifier for managing buckets. You can set tags for buckets, making it easy to group and manage them. For details, see set bucket tags.
Server-Side Encryption: You can choose the server encryption method. For the server-side encryption overview and supported regions, see server-side encryption overview.
4.3 Confirm
Confirm the bucket configuration information. If modification is needed, click Back.
5. After confirming the information is correct, click Create to create a bucket. In the bucket list interface, you can see the newly created bucket.

Note:
The bucket list only shows buckets created by Data Lake Compute (DLC). Configuration is not currently supported. To configure, go to the DLC console. For details, see managed storage configuration.

Using tools

You can use tools to create a bucket, such as the COSBrowser tool, COSCMD tool, etc. For more information, see Tools Overview.

Using the REST API

Use the REST API to initiate a bucket creating request. For more information, see PUT Bucket.

Using SDKs

Directly call the bucket creating method in the SDK. For more information, see the SDK documentation for the corresponding programming language below:
Android SDK
C SDK
C++ SDK
.NET SDK
Go SDK
iOS SDK
Java SDK
Node.js SDK
PHP SDK
Python SDK
Mini Program SDK


Querying Bucket

Last updated：2025-11-28 18:05:40

Overview

You can quickly query the created buckets by bucket name and tag in the COS console.
Note:
Before you can access the bucket list with a sub-account, the sub-account must be authorized by the root account. For more information, see Accessing Bucket List Using Sub-Account.
As the List Bucket operation is not restricted by bucket permission, a sub-account cannot be restricted to querying just specified buckets.

Usage

Using the COS Console

1. Log in to the COS console and click Bucket List on the left sidebar to enter the bucket list page, where you can view all the created buckets.
2. If there are many buckets, you can search for buckets by bucket name or tag in the top-right corner of the Bucket List page.
Query by bucket name: You can enter a specific bucket name or a bucket name prefix to query buckets.


Query by tag: If you have set tags for a bucket as instructed in Setting Bucket Tags, you can also select Tag in the top-right corner of the Bucket List page and enter a tag key to query buckets.



Using REST API/SDK

You can directly use the REST API to trigger a bucket query request. For details, see GET Service (List Buckets) API document.
You can directly call the SDK to create a bucket. For details, see the following SDK documentation for different languages:
Android, C++, .NET (C#), Flutter, Go, iOS, Java, Node.js, PHP, Python, React Native, Mini Program.

Using Tools

You can use tools to query buckets, such as COSBrowser, COSCLI. For more tool introduction, please see Tool Overview.

Clearing Bucket

Last updated：2025-09-02 11:13:54

Overview

You can clear the specified bucket in the COS console. For more information on buckets, see Bucket Overview.
Note:
Clearing a bucket will delete all objects and incomplete multipart uploads in this bucket. The deleted data is unrecoverable and inaccessible. Proceed with caution.

Use Cases

Clearing a bucket through a lifecycle rule: This method is applicable to a bucket with more than 10,000 objects. A deletion job will be triggered when the trigger condition of a lifecycle rule is met. The job start and completion time are subject to the lifecycle rule configuration in the console.
Quickly clearing a bucket in the console: This method is applicable to a bucket with less than 10,000 objects. A bucket clearing job will take effect immediately upon completion.
Note:
If you have a large amount of data in your bucket, clearing the bucket using the console may be slow or even fail due to network reasons. In this case, we recommend you clear the bucket by setting lifecycle configuration.

Directions

Quickly clearing a bucket in the console

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Find the target bucket and click More > Clear Data on the right.
4. In the pop-up window, enter the name of the bucket you want to clear and click Confirm.
5. In the pop-up window, click OK.

Clearing a bucket through a lifecycle rule

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Locate the bucket for which you want to set the lifecycle feature. Click the bucket name to enter its details page.
4. Click Basic Configurations > Lifecycle and click Add a Rule.
5. In the Add Rule dialog box, set rules to delete all current version files and historical versions in the entire bucket. Configuration as follows:
Basic information:
Status: Enabled.
Rule name: Required. Enter your lifecycle rule name.
Policy type: Select "Last modification time". If you have enabled the allowlist feature for access tracking, this configuration item will be displayed. If this configuration item is not present, the default policy type is last modification time without requiring selection. For more information, see Setting lifecycle.
Application scope: Select entire bucket.
Exclusion range: Off.
Rule configuration:
Manage current version files: Set deletion 1 day after file modification time.
Manage files of noncurrent version: Set deletion 1 day after file modification time.
Delete fragments: Set deletion 1 day after fragment creation.
Note:
Since deleting bucket data is a high-risk operation, secondary confirmation is required. Click OK to proceed.
6. You can see that the deletion rule has been set successfully in the lifecycle rule list.
7. After the lifecycle rule is executed, you can see that both the number and volume of stored objects are zero on the bucket overview page, indicating that the bucket has been cleared. If you no longer use the bucket, you can delete it.

Deleting a Bucket

Last updated：2025-09-29 10:07:05

When you no longer need to keep a certain Bucket, you can delete it. You can delete a COS Bucket on the Bucket list page through the COS console. You can also delete a Bucket using the console, tools, APIs, or SDKs.
Note:
A bucket cannot be recovered once deleted.

Prerequisites

Before executing bucket deletion, please confirm the following conditions must be met.
Completed the deletion of all files and their earlier versions in the bucket. For how to delete files and their earlier versions, see Clear the bucket.
Completed the deletion of all fragment files in the bucket. For how to delete fragments, see Delete incomplete multipart.
Closed CDN accelerated domain name, custom CDN acceleration domain name. To disable it, see CDN acceleration configuration.
Confirmed to have operation permissions for deleting bucket.
The root account has by default the permission to delete bucket. If your account is a sub-account, choose the following solution to obtain the permission to delete bucket:
Option 1: Add a favorite path: Use the root account to grant the sub-account specific bucket user permissions. After logging in with a sub-account to the COS console, add the bucket's favorite path to achieve bucket deletion operation permissions.
Option 2: Add preset policy: Add preset policy QcloudCOSFullAccess or QcloudCOSDeleteBucketAccess (permission to delete designated bucket) for sub-account to access bucket and implement bucket deletion operation.

Notes

Be sure to carefully read the following information, as ignoring it may cause disruption to your business.
When deleting Bucket, you need to ensure the identity is authorized for the operation and confirm the correct Bucket name and Region parameter are passed in.
After deleting Bucket, data cannot be recovered. Please confirm the data is no longer in use before deleting. If you need to continue using data in Bucket, please refer to Cloud Data Backup to plan in advance.
After deleting Bucket, you have to wait several hours (typically 4 to 8 hours) to create one with the same name. During this period, this name is open to everyone and can be used to create Bucket. If you hope to continue using the original Bucket name, it is advisable to empty content rather than delete directly so that Bucket is retained.

How to Use

Using COS console

1. Log in to the COS console. Click Bucket List on the left sidebar to open the bucket list page.
2. Find the target bucket and click More > Delete in the Operation column on the right.



3. In the pop-up window, perform the operation based on the system check result.

Using tools

Use tools such as COSBrowser and COSCMD to delete a bucket. For more information, see Tool Overview.

Using the REST API

Use the RESTful API to initiate a bucket deleting request. For more information, see DELETE Bucket in API documentation.

Using SDKs

Directly call the bucket deleting method in the SDK. For more information, see the SDK documentation for the corresponding programming language below:
Android SDK
C SDK
C++ SDK
.NET SDK
Go SDK
iOS SDK
Java SDK
JavaScript SDK
Node.js SDK
PHP SDK
Python SDK
Mini Program SDK

FAQs

Does it support batch deletion of Bucket?
Does it support deleting files of specified date in Bucket?
Version control is enabled, causing unable to clear the bucket or incomplete deletion. How to handle?
Will an empty bucket no longer uploading files cause charges?
Will deleting a bucket cause charges?
Unable to delete bucket due to arrears. What should be done?
Unable to cancel account. Resource detection prompts deleting bucket. How to handle?

Setting Hotlink Protection

Last updated：2025-11-28 17:49:13

Overview

Tencent Cloud COS provides hotlink protection to help users avoid unnecessary losses caused by malicious programs' stealing public network traffic using resource URLs or stealing resources. We recommend that you configure the blocklist/allowlist in Hotlink Protection Settings on the console to ensure security protection.
Note:
If a signature is carried in the access URL or headers, hotlink protection−based verification will not be performed.
When configuring hotlink protection, you can add your domain to the allowlist for the multipart upload request of large files.

Directions

1. Log in to the COS console. Click Bucket List on the left sidebar to open the Bucket List page.
2. Locate the bucket for which you want to set hotlink protection, and click its name to enter the bucket management page.
3. Click Security Management > Hotlink Protection, find the hotlink protection configuration item, and click Edit.
4. Switch the status to Enable, select a list type (blocklist or allowlist), input the domain name or IP address to be added to the blocklist or allowlist in the Referer configuration item, confirm and click save, configuration items are described as follows:



Blocklist: Select this item to deny access to the default access address of the bucket for domain names or IP addresses in Referer, and return a 403 error.
Allowlist: Select this item to allow access to the default access address of the bucket for domain names or IP addresses in Referer. Other domains or IPs will be rejected when accessing the default access address of the bucket, and a 403 error will be returned.
Empty referer: For an HTTP request, the header referer can be left empty (i.e., the HTTP request header has no referer field or the referer field is empty).
Referer: Enter up to 30 domain names or IP addresses (one per line). The wildcard * is supported, such as *.test.com. Examples are as follows:
If www.example.com is specified, www.example.com/123, www.example.com.cn, and other addresses with the prefix of www.example.com will also be included in the list.
Domain names and IPs with ports are supported, such as www.example.com:8080 and 10.10.10.10:8080.
If *.example.com is specified, addresses such as a.b.example.com/123 and a.example.com are also included.
Note:
If accelerated access is implemented via a CDN endpoint domain name, CDN hotlink protection rules will be executed before COS hotlink protection rules.

Using REST API/SDK

You can set hotlink protection through the following API:
PUT Bucket referer
GET Bucket referer
You can directly call the SDK to set hotlink protection. For details, please refer to the following SDK documentation for different languages:
Android, C, C++, .NET(C#), Go, iOS, Java, JavaScript, Node.js, PHP, Python, Mini Program.

Example

A user with the APPID of 1250000000 creates a bucket named examplebucket-1250000000 and places an image picture.jpg in the root directory. COS has generated the following default access address according to the rules:
examplebucket-1250000000.file.myqcloud.com/picture.jpg
User A owns a website:
www.example.com
and embeds the image into the homepage index.html.
Webmaster B manages a website:
www.fake.com
and wants to put this image on www.fake.com. But he doesn't want to pay for traffic costs. He creates a direct link to picture.jpg through the following address and places it into the homepage index.html on www.fake.com.
examplebucket-1250000000.file.myqcloud.com/picture.jpg
In such cases, to avoid losses for User A, we provide the following two methods of enabling hotlink protection.

Method 1

Configure the blacklist by entering the domain name *.fake.com, and save.

Method 2

The allowlist method: Add *.example.com to the allowlist and save.
The image is displayed normally when http://www.example.com/index.html is accessed. The image is also displayed normally when http://www.fake.com/index.html is accessed.
The image is displayed normally when http://www.example.com/index.html is accessed. The image cannot be displayed when http://www.fake.com/index.html is accessed.

Notes for Mini Program

1. For network requests using Weixin Mini Programs, the referer value is fixed as https://servicewechat.com/{appid}/{version}/page-frame.html.
2. If hotlink protection is enabled for your bucket, to allow Weixin Mini Programs to load COS images, add servicewechat.com to your hotlink allowlist in the COS console.

Setting Object Lock

Last updated：2024-01-06 14:59:34

Overview

COS's object lock feature allows you to set a retention period for your objects to prevent them from being overwritten or deleted for a fixed amount of time and your objects can still be accessed immediately. This document describes how to enable object lock in the COS console.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the target bucket to enter the bucket list page.
4. Click Security Management > Object Lock on the left, find the Object Lock configuration item, click Edit, and toggle the feature on.
5. Enter the retention period in the configuration window and click Save.

Retention period: It must be a positive integer and can only be extended but not shortened.
6. In the pop-up window, click OK. After the configuration, you can click File List > the desired file > Details to view the date (local time) on which the object lock rule will expire.
Note:
The object lock feature is now only available to customers in the allowlist. To use this feature, contact us.

Setting Origin-Pull

Last updated：2025-09-30 10:22:23

Overview

You can set origin-pull rules for buckets through the Cloud Object Storage (COS) console. When the object requested does not exist in the bucket or redirection of specific requests is needed, you can access the corresponding data from COS via origin-pull rules. Setting origin-pull is primarily used for hot migration of data, redirection of specific requests, and other scenarios. You can configure it according to your actual needs.
Note:
The success rate of data origin-pull depends on your network environment.Please give priority to using IP segments such as China Telecom, China Mobile, and China Unicom.
Buckets in the Financial Cloud region do not support origin-pull settings.
Public cloud regions only partially support synchronous origin-pull, including: Beijing, Chengdu, Guangzhou, Nanjing, Shanghai, Hong Kong (China), Singapore, Jakarta, Tokyo, Frankfurt, Virginia (US East), and Silicon Valley (US West).




Origin-Pull Rules

Trigger condition

In async and sync origin-pull modes, trigger origin-pull only when a GET request returns 404. If it is a HEAD request or other requests, it will not trigger origin-pull.
In redirect mode, you can customize HTTP status codes between 400 and 599 to trigger origin-pull.

Origin access

In async and sync origin-pull modes, you can set whether to pass through the QueryString and header information of the COS access request to the origin, and configure to carry additional header information when requesting the origin. In redirect mode, you can only set whether to pass through the QueryString.
If GET range is specified during the GET operation, COS will send an async request without range in addition to the original request to get the complete object data and store it in COS.

Response and storage

The origin can return data in chunked encoding.
If the origin returns a 404 status code, it will be passed through to COS and returned to the user. If 3xx Following Policy is enabled, when the origin returns a 3xx status code, COS will pull data from another origin. If the origin returns other non-3XX status codes, COS will return a 424 status code.
The file returned by origin-pull will be stored in COS with the filename used when the origin is requested. For example, if the file example.jpg requested by a user does not exist in the bucket, COS will trigger the origin-pull mechanism to pull the file from the configured origin-pull address http://origin.com/example.jpg and rename the file stored in the bucket example.jpg.
The new object stored in COS will contain the following metadata, with the data content following the values in the origin:
cache-control
content-disposition
content-encoding
content-type
expires
x-cos-meta-*

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket for which you want to set origin-pull.
4. Select Basic Configurations > Origin-Pull on the left, and click Add Rule.


5. In the pop-up window, configure the following information and click Next:

Origin-Pull Mode: Select an origin-pull mode as needed.
Asynchronous Origin-Pull: If the requested file doesn't exist in COS, COS will search for it on the specified origin. In the Async Origin-Pull mode, COS returns 302 to the client after finding that the file does not exist. The client jumps to the specified origin, and then asynchronously uploads the file to the COS bucket.
Note:
Async origin-pull does not directly return the file; instead, it returns a 302 status code to the client first and then asynchronously uploads the file to COS.
We recommend you enable Follow 302 to pull data from the origin.
The file upload time is subject to multiple factors, and no SLA can be promised. We recommend you select sync origin-pull if your business is sensitive to delays.
Sync Origin-Pull: If the requested file doesn't exist in COS, COS will search for it on the specified origin, return it to the client, and upload it to the bucket.
Note:
When using synchronous origin-pull, the timeliness can only be guaranteed if the file is fully downloaded and a status code 200 is returned. If the client accesses the file via a range request, the timeliness cannot be guaranteed.
Redirect: If a specified error is reported when accessing the bucket, COS will return the redirection address to the client but will not save the file from the origin. The client can request the resource from the origin at the redirection address.
Origin-Pull Condition:Select the conditions to trigger origin-pull as needed. Origin-pull is triggered only when all configured origin-pull conditions are satisfied simultaneously.
HTTP Status Code 404: If you select Async Origin-Pull or Sync Origin-Pull, trigger origin-pull when a GET request returns an HTTP status code of 404. This item is required and cannot be canceled. If you select Redirect, you can enter an HTTP status code ranging from 400 to 599.
File name prefix: When the requested file matches the prefix, the origin-pull rule will be triggered. For example, if you set this option to prefix, make a GET request to https://examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/prefix123.jpg and the returned HTTP status code is 404, the origin-pull rule will be triggered.
Origin-Pull Protocol: The protocol used by COS to access the specified origin. The options include Force HTTPS, Force HTTP, and Follow request protocol.
If you select Force HTTPS or Force HTTP, COS will access the origin using HTTPS or HTTP respectively.
If you select Follow request protocol, COS will access the origin with the protocol used in the request.
Request Parameter: Specify whether to pass through the queryString request parameters carried when accessing COS to the origin.
Request header: You can choose to pass through all request headers or pass through specified request headers. If you choose to pass through all request headers, it is recommended that you configure it to prohibit pass-through of the host header. One-click configuration is supported here.
Passthrough specifies the request header: When selecting pass through specified request header, you need to set this item. Here, you can add the header information that you want to be passed through to the origin server. If you select the redirection origin-pull type, do not set this item.
Prohibit passing through specified request header: Here you can add the header information in the request that you do not want to be passed through to the origin server. If you select the redirection origin-pull type, do not set this item.
New request header: Additional request headers can be added here when returning to the origin server. If the redirect back to the origin type is selected, this item should not be set.
6. Configure the following information based on the selected origin-pull mode and click Next:
Async origin-pull



Origin-Pull Address: Enter the domain name or IP address without the http:// or https:// prefix. You can also add the port number after the domain name or IP address.
Example of a correct address:
abc.example.com
abc.example.com:8080
202.96.128.86
202.96.128.86:8080
Standby forwarding address: Support setting the backup source address. For input format, refer to the configuration item description of origin address. This configuration depends on the bucket location. If the region supports it, it can be configured; otherwise, it cannot be configured.
Address settings: You can choose whether to set a specific file or file path to jump to. The configuration items are described as follows:
Fixed file settings: When the origin-pull rules are triggered, all requests will be redirected to the fixed file.
PrefixSuffixSettingsConfiguration: When the origin-pull rules are triggered, it will redirect to the file path with the specified prefix and suffix.
For example, if the prefix is set to test, accessing https://examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/path/prefix123.jpg triggers the origin-pull rules, and it will redirect to <origin address>/test/path/prefix123.jpg.
For example, if the suffix is set to .jpg, accessing https://examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/path/prefix123 triggers the origin-pull rules, and it will redirect to <origin address>/path/prefix123.jpg.
Note:
The configurations of specifying a prefix and specifying a suffix can take effect at the same time.
3xx following policy: Once enabled, when the origin server returns a 3XX redirection status code, COS's default handling policy is to follow the 3XX and fetch data from another origin server again. Selecting Disable will prevent resource retrieval.
Note:
When the origin server returns a 3XX redirection to an intranet address, COS will not follow it. In synchronous origin-pull mode, COS will not follow the redirected intranet address. In asynchronous origin-pull and redirection modes, after an intranet address is returned, business judgment is required to decide whether to follow or disable the redirection.
Sync origin-pull



Origin-Pull Address: Enter the domain name or IP address without the http:// or https:// prefix. You can also add the port number after the domain name or IP address. Example of a correct address:
abc.example.com
abc.example.com:8080
202.96.128.86
202.96.128.86:8080
Standby forwarding address: Support setting the backup source address. For input format, refer to the configuration item description of origin address. This configuration depends on the bucket location. If the region supports it, it can be configured; otherwise, it cannot be configured.
Address settings: You can choose whether to set a specific file or file path to jump to. The configuration items are described as follows:
Fixed file settings: When the origin-pull rules are triggered, all requests will be redirected to the fixed file.
PrefixSuffixSettingsConfiguration: When the origin-pull rules are triggered, it will redirect to the file path with the specified prefix and suffix.
For example, if the prefix is set to test, accessing https://examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/path/prefix123.jpg triggers the origin-pull rules, and it will redirect to <origin address>/test/path/prefix123.jpg.
For example, if the suffix is set to .jpg, accessing https://examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/path/prefix123 triggers the origin-pull rules, and it will redirect to <origin address>/path/prefix123.jpg.
Note:
The configurations of specifying a prefix and specifying a suffix can take effect at the same time.
3xx following policy: Once enabled, when the origin server returns a 3XX redirection status code, COS's default handling policy is to follow the 3XX and fetch data from another origin server again. Selecting Disable will prevent resource retrieval.
Packet of origin site: After this feature is enabled, the packet from the origin will be directly returned, including the status code and other information.This configuration depends on the bucket location. If the region supports it, it can be configured; otherwise, it cannot be configured.
Redirect



Origin-Pull Address: Enter the domain name or IP address without the http:// or https:// prefix. You can also add the port number after the domain name or IP address. Example of a correct address:
abc.example.com
abc.example.com:8080
202.96.128.86
202.96.128.86:8080
Address settings: You can choose whether to set a specific file or file path to jump to. The configuration items are described as follows:
Fixed file settings: When the origin-pull rules are triggered, all requests will be redirected to the fixed file.
PrefixSuffixSettingsConfiguration: When the origin-pull rules are triggered, it will redirect to the file path with the specified prefix and suffix.
For example, if the prefix is set to test, accessing https://examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/path/prefix123.jpg triggers the origin-pull rules, and it will redirect to <origin address>/test/path/prefix123.jpg.
For example, if the suffix is set to .jpg, accessing https://examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/path/prefix123 triggers the origin-pull rules, and it will redirect to <origin address>/path/prefix123.jpg.
Note:
The configurations of specifying a prefix and specifying a suffix can take effect at the same time.
Redirect Code: 301, 302, and 307 are optional. The default is 302.
7. Confirm that the configured origin-pull rule is correct and click OK.
8. By default, COS always gives the highest priority to the most recent rule, by which it performs origin-pull. To change the priority manually, you can click the Edit icon under the Priority column in the rule list.
Note：
After a high-priority source fails, it will not revert to a low-priority source address, but instead will revert to the backup source address of the original source.

Examples

Background A user whose APPID is 1250000000 created a bucket named "examplebucket-1250000000", and enabled CDN acceleration endpoint domain name:
examplebucket-1250000000.file.myqcloud.com
Configure the origin-pull address of the bucket to be:
abc.example.com
Store the image picture.jpg on the origin http://abc.example.com.
First access by client (without sync origin-pull enabled):
http://examplebucket-1250000000.file.myqcloud.com/picture.jpg
When COS finds that the object cannot be hit, it returns HTTP status code 302 to the client and redirects to the following address:
http://abc.example.com/picture.jpg
First access by client (with sync origin-pull enabled):
http://examplebucket-1250000000.file.myqcloud.com/picture.jpg
When COS finds that the object cannot be hit, it returns the HTTP status code 200 to the client and redirects to the following address:
http://abc.example.com/picture.jpg
The origin then provides the object to the client to ensure access, and COS copies picture.jpg from the origin and saves it to the root directory of the bucket "examplebucket-1250000000".
Second-time access:
http://examplebucket-1250000000.file.myqcloud.com/picture.jpg
COS directly hits the picture.jpg object in the root directory and returns it to the client.


Setting Cross-Origin Resource Sharing (CORS)

Last updated：2025-08-27 11:38:10

Overview

You can set Cross Origin Resource Sharing (CORS) for objects in buckets through the Cloud Object Storage (COS) console. COS supports configuring multiple rules to respond to OPTIONS requests. CORS is a mechanism that allows resources at one origin to be requested from another origin through HTTP requests. Origins are deemed different from each other as long as their protocols, domain names, or ports are different.
COS supports response to OPTIONS requests for CORS, and returns specific rules set by developers to browsers, but the server does not verify whether subsequent cross-origin requests conform to the rules. For more information, please see Cross-Origin Resource Sharing and Setting Cross-Origin Access.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Locate the bucket for which you want to set CORS and click the bucket name to enter the bucket details page.
4. Select Security Management > CORS (Cross-Origin Resource Sharing) on the left sidebar and click Add a Rule.



5. Add rule information (Fields with * are required). Configuration items are as follows:

Origin: The sources from which cross-origin requests are allowed. Add your own domain name. Supports adding domain names (including Chinese domain names) and IP addresses.
The domain name does not need to end with a slash (/).
More than one domain name can be specified, with one domain name per line.
You can set a *, indicating that all domains and IPs are supported, which is not recommended.
Note:
If the user has enabled xhr.withCredentials = true for the AJAX request (indicating that the request needs to carry domain cookies), then in this scenario, configuring Access-Control-Allow-Origin: * is not allowed; a specific domain must be configured. However, COS domains are generally used for static websites and do not require cookies, so it is recommended not to enable xhr.withCredentials = true.
A single specific domain name is supported, such as http://www.abc.com.
Second-level wildcard domain names, for example, http://*.abc.com, are supported. Note that each line can contain only one asterisk (*). If configured as http://*.example.com, it allows sources such as http://a.b.example.com and http://a.example.com.
Be careful not to omit the protocol name (http or https). If the port is not the default port 80, you need to add a port. An example IP address is http://10.10.10.10.
Chinese domain name is recommended to add both Chinese domain name and punycode domain name. For exampleChinese.cn and https://xn--fiq228c.cn.
Allow-Methods: GET, PUT, POST, DELETE, and HEAD are supported. Enumeration of one or more methods is allowed for a cross-origin request.
Allow-Headers: notifies the server about which custom HTTP request headers are allowed for subsequent requests when an OPTIONS request is sent, such as x-cos-meta-md5.
More than one header can be specified, with one header per line, for example, Content-Type.
Header is easy to omit. Therefore, if there is no special requirement, you are advised to set this field to *, meaning that all headers are allowed.
Uppercase and lowercase letters [a-z, A-Z] are supported.
Each header specified in Access-Control-Request-Headers must correspond to a value in Allow-Headers.
Expose-Headers: Request headers commonly used in COS. For more information, please see Common Request Headers. Set this parameter according to the specific application requirements. ETag is recommended. Headers do not allow wildcards and are case insensitive. You can set multiple headers, with one header per line.
Max-age: validity period (in seconds) of the results obtained by OPTIONS. The value must be a positive integer, such as 600.
Return Vary: Origin: Set whether to return the Vary: Origin header. Enable this option if the browser has both CORS and non-CORS requests; otherwise, cross-origin access issues will occur.
Note:
Enabling the Vary: Origin configuration may increase browser cache or CDN origin fetches.
Vary: Origin This is a global configuration item. After modification, the status of this field in all cross-origin rules will be updated as well.
6. After configuration, click Save and you will see the CORS rule added. To modify it, click Edit.




Accessing Bucket List Using Sub-Account

Last updated：2025-11-11 17:47:21

Overview

Sub-accounts have no permissions to pull a bucket list by default. Therefore, if you log in to the COS console with a sub-account, you cannot access overview data, bucket list, or any other administration items that require permissions.
Sub-accounts can access a bucket list using the following two methods:
Adding an access path: this method applies to scenarios where a sub-account has permissions to operate on objects but no permissions to access a bucket list. The access path can be a Bucket or a Path under the bucket. Please make sure the added path is authorized.
Adding a preset policy: you can add a preset policy QcloudCOSGetServiceAccess with your root account for a sub-account to access a bucket list. This method also allows you to check the statistics overview in the console.
Note:
This feature applies to scenarios where a sub-account accesses a bucket list using the console.

Adding an Access Path

Sub-accounts are not granted the preset policy QcloudCOSGetServiceAccess by default and thus do not have the permission to pull the bucket list. When granted the permissions (e.g., Read or Write) to a bucket by the root account, a sub-account can then access this bucket by adding an access path.

Prerequisites

The sub-account has been granted user permissions on a bucket by the root account. For more information, see Set Access Permission.

Directions

1. Log in to the COS console with a sub-account, enter the Access Path List page, and click Add Access Path.
2. In the Add Access Path pop-up window, select the bucket region and enter the access path, as shown below:
Region: select the region of the bucket to be allowed for access.
Access Path: enter the name of the bucket to be allowed for access (e.g., examplebucket-1250000000) or the path to an object in the bucket (e.g., examplebucket-1250000000/doc/exampleobject.txt).
3. After confirming that the region and the access path are correct, click OK to add the path to the authorized bucket or an object in it.
4. Click Objects on the right, and you can see the object(s) to which the sub-account has been granted access.

Adding a Preset Policy

A sub-account can access the bucket list by adding the preset policy QcloudCOSGetServiceAccess (i.e., the permission to obtain the bucket list) to it.
Note:
The preset policy QcloudCOSFullAccess or QcloudCOSReadOnlyAccess can also grant a sub-account access permission to the bucket list. However, due to the wide coverage of permissions granted by these two policies, they are not recommended for security reasons.
The collection of statistics in the overview requires the access permission to the bucket list. When the sub-account needs to pull statistics, please make sure that the root account has added the preset policy QcloudCOSGetServiceAccess to it; otherwise, the system will prompt that the sub-account has no access permission to the statistics.
1. Log in to the CAM console with the root account to enter the user list page.
2. Locate the sub-account to which you want to add a policy, and click Grant Permission on the right to enter the Associate Policies page.
3. Search for and add the preset policy QcloudCOSGetServiceAccess (i.e., the permission to access the bucket list in COS) in the policy list.
4. Click OK.
5. Click the sub-account name to enter its details page where you view the added policies. When you no longer need a policy, you can unbind it.
Note:
Now, you have successfully added a preset policy for the sub-account through the root account. Log in to the COS console with the sub-account, and you can check the bucket list and statistics overview.

Domain Name Management

Overview

Last updated：2024-01-06 14:59:34

Overview

After an object is uploaded to the bucket, COS will automatically generate a URL (i.e., default domain name) for you to access this object directly. To use CDN or your domain name to access COS objects, you can bind your domain or CDN to the bucket where the objects are stored.
You can set a domain name to access objects as needed. To accelerate access using CDN, you can access using the URL generated with the CDN acceleration domain name.

Description

You can quickly download and deliver objects in a bucket by managing the following domain names:
Default domain name: COS origin domain name, which is automatically generated based on the bucket name and region when you create a bucket.
Custom CDN acceleration domain name: you can bind for your bucket a custom domain name to Tencent Cloud Content Delivery Network (CDN), and access objects in your bucket using this domain name. (If you have enabled "User-defined domain name" in the legacy COS console, then the new console will continue to display "User-defined domain name" instead of “Custom CDN acceleration domain”.)
Custom endpoint: Allows you to bind your own domain name as a custom endpoint to the bucket for access to the objects in it.
Note:
Currently, you must activate the CDN service to use a custom domain name in COS.
For content delivery in the Chinese mainland, ICP filing is required. You are not required to do so through Tencent Cloud though.
For content delivery outside the Chinese mainland, ICP filing is not required, but note that your data and operations in Tencent Cloud still need to comply with local laws and regulations as well as General Service Level Agreements.
With CDN acceleration enabled for the custom CDN acceleration domain name, if the origin is a public-read bucket, the objects in the origin can be accessed via the custom CDN acceleration domain name. If the origin is a private-read bucket, we recommend you enable the CDN origin-pull authentication and CDN authentication configuration options.
Origin-pull authentication (CDN service authorization must be added before it can be enabled): If the data requested by a user is not cached in the edge node, CDN fetches the data from the origin. If COS is used as the origin and origin-pull authentication is enabled, the CDN edge node accesses the COS origin using a special service identity (which must be authorized by CDN service) to acquire and cache the data in the private bucket.
CDN authentication: When a user accesses an edge node to acquire cached data, the edge node verifies the authentication field in the access URL based on the authentication configuration rules. This prevents unauthorized access and hotlinking, thereby improving the security and reliability of the data cached in the edge node.
CDN authentication configuration and CDN origin-pull authentication do not conflict with each other, but whether to enable them can affect the level of data protection, as shown below:
Bucket access permission
CDN origin-pull authentication
CDN authentication
Origin can be accessed via CDN acceleration domain name
Origin can be accessed via COS endpoint
Scenarios
Public read
Disabled
Disabled
Yes
Yes
Site-wide public access
Public read
Enabled
Disabled
Yes
Yes
Not recommended
Public read
Disabled
Enabled
URL authentication is required
Yes
Not recommended
Public read
Enabled
Enabled
URL authentication is required
Yes
Not recommended
Private read + CDN service authorization
Enabled
Enabled
URL authentication required
COS authentication required
Full-linkage protection
Private read + CDN service authorization
Disabled
Enabled
URL authentication is required
COS authentication is required
Not recommended
Private read + CDN service authorization
Enabled
Disabled
Yes
COS authentication is required
Origin protection
Private read + CDN service authorization
Disabled
Disabled
No
COS authentication is required
Not recommended
Private read
Disabled
Enabled or disabled
No
COS authentication is required
CDN is unavailable
Note:
Take the first row in the above list as an example. If the origin bucket is public read, and neither CDN origin-pull authentication nor CDN authentication configuration is enabled, you can access CDN edge nodes and the origin bucket using the CDN domain name, and access the origin bucket using the COS domain name.
Origin protection is useful in cases where your data cached on CDN edge nodes may be maliciously pulled due to a lack of CDN authentication. Therefore, we strongly recommend you enable CDN authentication to mitigate data security issues.
After CDN acceleration is enabled for a domain name, anyone can directly access the origin via the domain name. Therefore, if you need to keep your data private, be sure to protect your data in the origin through Authentication Configuration.

More

Enabling Custom Acceleration Domain Names
Enabling Custom Origin Domain Names
Granting a sub-account permissions to configure bucket acceleration domain names
Supporting HTTPS for Custom Endpoints

Enabling Custom CDN Acceleration Domain Name

Last updated：2025-03-24 14:31:48

Overview

This document only describes how to add a custom acceleration domain name and enable CDN acceleration in the COS console. For more information on how to add a custom domain name by using the CDN console, see Adding Domain Names.

Directions

1. Log in to the COS console. Click Bucket List on the left sidebar to open the Bucket List page.
2. Click the bucket you need to set a domain name to go to the bucket configuration page.
3. Select Domains and Transfer > Custom CDN Acceleration Domain, in the Custom CDN Acceleration Domain area, click Add Domain, and configure the following options.
Note:
If you have enabled Custom Domain in the previous console, you will see Custom Domain still displayed, with Custom CDN Acceleration Domain not displayed in the current console.
Domain Name: Enter the target custom domain name (such as www.example.com). Make sure that an ICP filing has been obtained and a CNAME record has been configured at the DNS service provider for the entered domain. For more information, see CNAME Configuration. If the custom CDN acceleration domain you are connecting is in the following situations, you need to verify your domain ownership as instructed in Domain Name Ownership Verification.
The domain name is being connected for the first time.
The domain name has been connected by another user.
The domain name is a wildcard domain name.
Acceleration Region: CDN acceleration is supported for regions in the Chinese mainland, outside the Chinese mainland, and globally, with global acceleration for buckets across all regions.
Origin Server Type: Support Default Endpoint, Static Website Endpoint, and Global Acceleration Endpoint. If your bucket has enabled Static Website or Global Acceleration, support selecting Static Website Endpoint or Global Acceleration Endpoint.
Authentication: Enable origin-pull authentication. For private-read buckets, enable Origin-pull Authentication to protect the origin server.
Note:
For private-read buckets, if both origin-pull authentication and CDN service authorization are enabled, signature is not required for accessing the origin via CDN, and cached resources in CDN will be distributed on the public network, which will affect the data security. Therefore, we recommend you enable CDN authentication (Step 5).
Automatic CDN Cache Purge: After this feature is enabled, when a file in the COS bucket is updated, the CDN cache will be automatically purged. You can go to the Function Service section in the COS console to configure this feature as instructed in Setting CDN Cache Purge.
HTTPS Certificate: To add an HTTPS certificate for the custom domain name, go to the CDN console.
4. After the configuration, click Save in the Operation column on the right to add the domain name. After it is saved, the Enable button for CDN authentication appears next to Authentication. You can click the button to enable the CDN authentication for the custom domain name. CDN Authentication: Timestamp authentication can be configured to prevent stealing by malicious users. You can enable the feature after adding the domain name.
Note:
After CDN acceleration is enabled for a domain name, anyone can directly access the origin via the domain name. Therefore, if you need to keep you data private, be sure to protect your data in the origin through Authentication Configuration.
5. Log in to the CDN console and click Domain Management on the left sidebar.
6. Locate the domain name you need to configure and click Manage under Operation on the right to open the domain management page. Click Access control at the top, and then locate Authentication Configuration. For details, see Instruction.

Enabling Custom Origin Server Domains

Last updated：2025-11-18 15:12:40

Overview

This document describes how to bind a custom domain name to a bucket. You can access the files in the bucket via the custom domain name.
Note:
Up to 20 custom domain names can be added via the COS console. If you want to raise the upper limit of the number of custom domain names, contact us.
After enabling the custom domain name, you also need to configure CNAME domain name resolution for the changes to take effect. For details, see Configure Domain Name Resolution.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the bucket for which you need to set a domain name to go to the bucket configuration page.
4. In the left sidebar, select Domains and Transfer > Custom Endpoint.
5. On the Custom Endpoint page, click Add Domain. Configuration items are described below.



Domain Name: Enter the custom domain name to be bound (for example, www.example.com) or scroll down to select an existing Tencent Cloud domain name. Please ensure that the entered domain name has been added for resolution at the DNS service provider.
Note:
If you have no available domain name (or need to add a new domain name), click Register a new domain to quickly go to the domain name registration console to perform registration.
After the new domain name is registered successfully, you can select "Tencent Cloud domain name" and drop down to select that domain name, then continue with the next operation.
CNAME: Configure the corresponding CNAME record at the DNS service provider to route the request to COS.
Origin Server Type: Divide into the following several types.
Default Endpoint: If you wish to use the custom domain name as the default endpoint, please select the default endpoint.
Static Website Endpoint: If you wish to use the custom domain name as a static website, first enable the static website functionality for the bucket, and then select the static website endpoint.
Global Acceleration Endpoint: If you wish to use the custom domain name for global acceleration, first enable the global acceleration feature for the bucket, and then select the global acceleration endpoint.
6. After the configuration is complete, click Create to see the configured custom endpoint domains on the list page.
Note:
Addition, launch, or offline of domain name operations do not take effect immediately. The actual configuration takes some time and generally takes effect 30 minutes later. It is subject to the actual access status of your domain.
7. If you need to configure a certificate, click
Bind Certificate
to configure. After binding, it is expected to take effect within 30 minutes. The certificate source supports selecting a newly uploaded certificate or a hosted certificate, as follows.
Newly uploaded certificate: The original self-owned certificate. If you wish to use a self-owned certificate, please select this item.
Hosted certificate: The original Tencent Cloud certificate. If you wish to use a certificate applied for by using Tencent Cloud, please select this item.
8. According to the selected certificate source, configure the following information.
Note:
COS in all regions has supported managed HTTPS certificates for custom endpoint domains. If your domain name does not yet have an HTTPS certificate, you can click Apply for Tencent Cloud certificate.
You can also achieve HTTPS access by configuring a custom domain name via reverse proxy (with CDN acceleration disabled). Refer to method 2.
Newly Uploaded Certificate
Certificate content：
Normally, a certificate is a file with an extension such as.crt or.pem (for example, imgsc.aa.com_bundle.crt). Please use a corresponding text editor to open the certificate file.
The certificate format begins with "-----BEGIN CERTIFICATE-----" and ends with "-----END CERTIFICATE-----".
Include the complete certificate chain in the certificate content.
private key content：
Normally, a private key is a file with an extension such as.key or.pem (for example, imgsc.aa.com.key). Please use a corresponding text editor to open the private key file.
The private key format begins with "-----BEGIN (RSA) PRIVATE KEY-----" and ends with "-----END (RSA) PRIVATE KEY-----".
Managed SSL:
Supports hosting newly uploaded certificates to SSL, with default check. It is recommended that you host them on SSL. After hosting, you can manage the certificates in the SSL console and enjoy managed services such as certificate expiration notifications, alarms, and updates.
Normally, a private key is a file with an extension such as.key or.pem (for example, imgsc.aa.com.key). Please use a corresponding text editor to open the private key file.
The private key format begins with "-----BEGIN (RSA) PRIVATE KEY-----" and ends with "-----END (RSA) PRIVATE KEY-----".
Hosted Certificate
Certificate list: Show the hosted certificates list. Include certificate ID, domain name, and certificate expiration time.
9. After successful binding, on the custom endpoint list page, you can see the certificate ID and expiration date of the current certificate.
10. If you wish to rebind the certificate, click Bind Certificate. If you do not wish to bind the current certificate, click Unbind Certificate.

Appendix: Configure Domain Name Resolution

Go to the appropriate DNS service to configure the CNAME Resolution Record. Directions:
1. Add domain: Manually input the custom endpoint domain to be resolved, for example, www.example.com.
2. Add a record: Add a resolution record with record type CNAME for the custom endpoint domain. The record value is the default endpoint of the bucket pointed to by the CNAME, for example, bucket-1250000000.cos.ap-beijing.myqcloud.com.
3. Verify whether the resolution is successful: The resolution record takes some time to take effect. You can use the dig command or COS console to view whether the resolution is successful. Follow these steps for verification:
Enter the command dig mydomain.com in the command prompt to check whether the CNAME record has taken effect correctly. (Replace mydomain.com with your custom domain name during use.)
Log in to the COS console, view the custom domain name for bucket. If the CNAME of the domain name does not take effect successfully, a corresponding notification will appear.

Granting Sub-Account Permission to Configure Bucket Acceleration Domain Names

Last updated：2024-01-06 14:59:34

Overview

On the Domain Management Configuration page of a COS bucket, you can configure the default CDN acceleration domain name, custom CDN acceleration domain name, and custom endpoint. Among them, the configurations of the first two are logically related to the CDN service. Therefore, if you want a sub-account to configure them, in addition to COS management permissions (to configure buckets, for example), you must also grant the sub-account relevant permissions of the CDN service.
To ensure the resource security, if you don't grant the sub-account relevant permissions of the CDN service, the sub-account will not be able to configure the default and custom CDN acceleration domain names for a COS bucket by default. If an unauthorized sub-account logs in to the COS console and navigates to the Domain Management Configuration page, an access denied error will be displayed.

If you want the sub-account to properly configure such domain names, you need to authorize it in the CAM console by following the steps below:

Directions

1. Log in to the CAM console and go to the Policies page.
2. Click Create Custom Policy > Create by Product Feature or Project Permission to enter the Configure Service Types page.
Note:
The permissions can be configured by the root account by default. If you are using a sub-account and want to configure the permissions here, make sure that the root account has granted you permissions to do so. The user policy that should be authorized by the root account in this case is QcloudCamFullAccess.
3. Enter your policy name (e.g., COS_DomainAccess), select CDN as the Service Type, and click Next.
4. Grant corresponding feature APIs to the sub-account based on your business needs. Access to and configuration of the default and custom CDN acceleration domain names for a COS bucket involve five features: querying domain name information, adding domain names, enabling/disabling domain names, deleting domain names, and modifying domain name configurations. If you want the sub-account to have full access to the configurations of all domain names on the Domain Management Configuration page, toggle on all these features.
5. After selecting the corresponding features, click Next to associate objects.
6. Associate features with objects. Select Associate Objects > All (including resource objects purchased in the future), which must be selected to make the policy configuration fully effective.
7. Confirm that the permissions are correctly configured and click Complete to create the custom policy.
8. After the custom policy is created, switch to the User List page, and then associate the sub-account with the policy by clicking Authorize on the right.

9. In the Associate Policy pop-up window, search for and select the custom policy you just created and click OK.

10. After the policy is associated, the sub-account is authorized and can log in to the COS console to access and configure the default and custom CDN acceleration domain names for a COS bucket.


Object

Object Overview

Last updated：2025-11-19 11:56:13

Overview

Object is the basic unit of storage in COS, which can be understood as data in any format, such as images, documents, and audio/video files. Bucket is the carrier of objects. One bucket can contain any number of objects. Different storage classes can be specified for objects in COS. For more information, see Storage Class Overview.
Each object consists of an object key (ObjectKey), an object value (Value), and object metadata (Metadata).
ObjectKey: The unique identifier of an object in a bucket, which can be regarded as the file path. In the API and SDK samples, objects are named in the format of <ObjectKey>.
Value: The uploaded object itself, which can be commonly understood as the object content.
Metadata: A set of key-value pairs, which can be regarded as file attributes, such as file modification time and storage class. You can query them after uploading the object.

Object Key

An object in COS must contain a valid ObjectKey as the unique identifier of the object in a bucket. For example, in an object's access address examplebucket-1250000000.cos.ap-guangzhou.myqcloud.com/folder/picture.jpg, the object key is folder/picture.jpg.

Naming conventions

You can use any UTF-8 characters in an object key name. However, we recommend you use letters and digits to ensure optimal compatibility with other applications.
The encoded length can be up to 850 bytes.
The name cannot start with a forward slash / or backslash \.
An object key cannot contain certain ASCII control characters, including upward arrow (↑), downward arrow (↓), rightward arrow (→), and leftward arrow (←), corresponding to CAN (24), EM (25), SUB (26), and ESC (27) respectively.
Refrain from using special symbols such as * and % in the filename.
Note:
If the name of the uploaded file or folder contains Chinese characters, when you access or request the file or folder, the Chinese characters will be converted into a percent-encoded string based on URL-encoding rules.
For example, when you access XXX.doc, the object key is XXX.doc, but the percent-encoded string actually read is %e6%96%87%e6%a1%a3.doc.
The following are examples of valid object key names:
doc/exampleobject
my.great_photos-2016/01/me.jpg
videos/2016/birthday/video.wmv

Special symbols

Certain characters may need to be URL-encoded or referenced in the hexadecimal format. Some of them are non-printable characters, and your browser may not be able to handle them. They require special handling as detailed below:
,
:
;
=
&
$
@
+
?
ASCII character ranges: 00–1F hexadecimal (0–31 decimal) and 7F (127 decimal)
(space)

There are also some characters that require significant special handling to maintain consistency across all applications, so we recommend you avoid them directly as detailed below:
`
^
"
\
{
}
[
]
~
%
#
|
*
>
<
ASCII 128–255 decimal


Object access address

The access address of an object consists of a bucket access address and an object key in the format of <bucket domain name>/<object key>. For example, after you upload the exampleobject.txt object to the examplebucket-1250000000 bucket in Guangzhou (South China) region, the access address for exampleobject.txt is examplebucket-1250000000.cos.ap-guangzhou.myqcloud.com/exampleobject.txt.

Folder and directory

As COS comes with no folders or directories, it will not create a project folder for uploading the object project/a.txt. To make it easier for you to get started, COS simulates the display mode of "folder" or "directory" in the console and graphical tools such as COSBrowser. This is implemented by creating an empty object with a key value of project/ and displaying it as a traditional folder.
For example, when you upload the object project/doc/a.txt via APIs or SDKs, the separator / simulates the display mode of "folder", and you can see the folders project and doc in the console. The folder doc is displayed under the folder project and contains the file a.txt.
Note:
Objects in the bucket are evenly distributed among distributed clusters. Therefore, you cannot directly get the size of all objects with a specified object key prefix. Instead, you can accumulate the size of each object to get the full size.
Deleting folders and directories is relatively complicated as detailed below:
Deletion Method
Object or Folder Deletion
Result
Console
Folder project
All objects with the object key prefix project/ will be deleted.
Console
Object project/doc/a.txt
The project and doc folders will be retained.
API and SDK
Object project/ or project/doc/
The object project/doc/a.txt will be retained. If you want to delete the objects in the folder, use code traversal to delete them.

Object metadata

Object metadata (or HTTP header) is a set of key-value pairs in an object. It is the string sent by the server before the server sends HTML data over the HTTP protocol to the browser. Modifying the HTTP header when uploading an object can alter page response forms or communicate configuration information, such as modifying caching time.
There are two types of metadata: system metadata and user-defined metadata.
Note:
Modifying an object's HTTP header does not modify the object itself.

System metadata

This refers to the attribute information of an object, such as object modification time, object size, and storage class.
Name
Description
Content-Length
The length of the content of an HTTP request in bytes as defined in RFC 2616
Last-Modified
The creation date or last modified date of an object (whichever is later)
x-cos-version-id
Object version ID. If versioning is enabled for a bucket, the version ID of the object will be returned to identify the historical versions of the object.
ETag
The MD5 value of an object uploaded by calling `PUT Object`, or the unique ID of an object uploaded by multipart upload or a legacy API (which cannot perform check though).


User-defined metadata

This refers to the object's custom parameters, such as Content-Type, Cache-Control, Expires, and x-cos-meta-*. For more information, see Custom Headers.
Name
Description
Cache-Control
The caching policy as defined in RFC 2616, which is saved as part of the object metadata.
Content-Disposition, Content-Encoding, Content-Type
The filename/encoding format/content type (MIME) as defined in RFC 2616, which are saved as part of the object metadata.
Expires
The cache expiration time as defined in RFC 2616, which is saved as part of the object metadata.
x-cos-acl
ACL attributes of an object. Valid values: `private`, `public-read-write`, `public-read`. Default value: `private`.
x-cos-grant-*
Grants the grantee permissions.
x-cos-meta-*
User-defined header, which is returned as part of the object metadata with a size up to 2 KB.
x-cos-storage-class
Storage class of an object. Enumerated values: `STANDARD`, `STANDARD_IA`, `ARCHIVE`. Default value: `STANDARD`.
x-cos-server-side-encryption
Specifies whether to enable server-side encryption for the object and the encryption method.

Object Operations

You can manage objects through the Tencent Cloud console, tools, APIs, and SDKs.
Note:
Objects can be uploaded via simple upload or multipart upload based on the object size.
Use simple upload for objects less than 5 GB in size.
Multipart upload is limited to no more than 10,000 parts (each in 5 GB) and a maximum object size of around 48.82 TB. For more information on use limits, see Specifications and Limits.
After uploading objects, you can configure them in the console. For more information, see:
Searching for Objects
Viewing Object Information
Setting Object Access Permission
Custom Headers

Object Subresource

COS has subresources that are associated with buckets and objects. Subresources are subordinates to objects, so they do not exist on their own. They are always associated with other entities such as objects or buckets. An ACL is the access control information list for a specific COS object, which is a subresource of the object.
An ACL contains an authorization list that identifies authorized users and the granted permissions to implement access control on the object. When an object is created, the ACL identifies the object owner who can fully control the object. You can search for the object ACL or replace it with a new one.
Note:
You can update an ACL only by replacing it.

Access Permission Types

COS supports setting two permissions for objects: public permissions and user permissions. Public permissions: Include Inherit, Private Read/Write, and Public Read/Private Write.
Inherit: The object inherits the access permission of the bucket. When you access an object with the inherited bucket permission, COS will match the bucket permission to respond to the access request. A new object inherits the permission from its bucket by default.
Private Read/Write: When you access an object with the private read/write permission, the object can only be accessed with a request signature, regardless of the bucket access permission.
Public Read/Private Write: When you access an object with the public read permission, the object can be directly downloaded, regardless of the bucket access permission.
User permissions: A root account has all the permissions (full access) of buckets by default. In addition, you can add sub-accounts and grant them permissions to read/write data and permissions and even get full access.

Use cases

Allow public access to a specified object in a private read/write bucket, or require authentication for a specific object in a public read/write bucket.

Storage Class

Storage Class Overview

Last updated：2025-10-20 15:21:50

Storage class reflects the storage level and activity level of objects in Cloud Object Storage (COS). Based on access frequency and disaster recovery level, COS provides multiple storage types: MAZ_STANDARD, MAZ_STANDARD_IA, MAZ_ARCHIVE, MAZ_INTELLIGENT TIERING, INTELLIGENT TIERING, STANDARD, STANDARD_IA, ARCHIVE, and DEEP_ARCHIVE. Each storage type has varying characteristics, such as object access frequency, data durability, data availability, and access delay. Users can choose the storage type to upload data to COS based on their scenario.
Note:
If you did not specify the storage class when uploading an object, it will be uploaded to STANDARD by default.
For more information on multi-AZ, see Overview of Multi-AZ Feature.

MAZ_STANDARD/STANDARD

Both MAZ_STANDARD and STANDARD storage classes are highly reliable, available, and powerful object storage service designed for hot data and feature low latency and high throughput.
MAZ_STANDARD has higher data durability and service availability than STANDARD. It uses a different storage mechanism to store the data in different data centers in the same region, so as to prevent failures in one data center from affecting the entire service and further guarantee your business stability.
Use cases
Use cases involving lots of hotspot files or frequent data access, including trending videos, social images, mobile apps, game programs, and static websites.
The STANDARD storage class is designed for general use and covers most use cases. It is more cost-effective than MAZ_STANDARD.
However, MAZ_STANDARD has higher data durability and service availability, making it suitable for business scenarios with higher requirements, including key files, commercial data, and sensitive information.

MAZ_STANDARD_IA/STANDARD_IA

Both MAZ_STANDARD_IA and STANDARD_IA storage classes are highly reliable object storage services with low storage costs and access latency. They allow you to access the first byte in milliseconds at a reduced price, so you can retrieve data quickly without waiting. Unlike STANDARD, they involve data retrieval fees when you access data.
MAZ_STANDARD_IA uses a different storage mechanism from STANDARD_IA to store the data in different data centers in the same region, so as to prevent failures in one data center from affecting the entire service and further guarantee your business stability.
Use cases
Use cases with low access frequency (for example, 1 to 2 times per month), such as cloud disk data, big data analysis, government and enterprise data, low-frequency archives, and monitoring data.
Note:
Both MAZ_STANDARD_IA and STANDARD_IA have minimum storage requirements. If the storage duration is less than 30 days, the bill is calculated as 30 days. Likewise, if the size of a file is smaller than 64 KB, the bill is calculated as 64 KB (if the size of a file is greater than or equal to 64 KB, the bill is calculated based on the actual file size). For more information, see Pricing | Cloud Object Storage.

MAZ_INTELLIGENT TIERING/INTELLIGENT TIERING

Objects in the MAZ_INTELLIGENT_TIERING storage class can be stored in two storage layers: MAZ_STANDARD and MAZ_STANDARD_IA. Objects in the INTELLIGENT_TIERING storage class can also be stored in two storage classes: STANDARD and STANDARD_IA. COS will automatically switch between storage classes based on the access frequency of such objects with no data retrieval fees incurred, which reduces your storage costs. For more information, see INTELLIGENT TIERING Overview.
MAZ_INTELLIGENT TIERING uses a different storage mechanism from INTELLIGENT TIERING to store the data in different data centers in the same region, so as to prevent failures in one data center from affecting the entire service and further guarantee your business stability.
Use cases
Use cases with uncertain data access patterns. If your business has tight controls on costs and is less sensitive to file reading performance, you can use MAZ_INTELLIGENT TIERING or INTELLIGENT TIERING to reduce costs.
Note:
For MAZ_INTELLIGENT_TIERING and INTELLIGENT_TIERING, objects are billed based on their actual sizes. For more information on pricing, see Pricing | Cloud Object Storage.

MAZ_ARCHIVE/ARCHIVE

MAZ_ARCHIVE and ARCHIVE belong to Cold Data Type, providing users with high reliability, extremely low storage cost, and long-term storage service. Both MAZ_ARCHIVE and ARCHIVE require a min 90-day storage time, and data recovery (unfreeze) is required before reading data.
COS supports the following three restoration modes for ARCHIVE:
Expedited: Restores an object within 1-5 minutes.
Standard: Restores an object within 3-5 hours.
Bulk: Restores objects within 5-12 hours.
MAZ_ARCHIVE uses a different storage mechanism compared with ARCHIVE , storing data in different data centers in the same city to further guarantee business stability against failure influence of a single IDC.
Note:
For more information about object restoration, see Restoring Archived Objects.
The QPS for restoration requests is limited to 100.
Use cases
MAZ_ARCHIVE and ARCHIVE are suitable for business scenarios requiring long-term data preservation, such as archival data, medical imaging, scientific materials, compliance file archiving, lifecycle file archiving, operation log archiving, and remote disaster recovery.
Note:
Both MAZ_ARCHIVE and ARCHIVE have minimum limits for storage time and storage unit. If the storage time is less than 90 days, it will be counted as 90 days. Each stored file smaller than 64KB will be counted as 64KB, while files greater than or equal to 64KB will be charged based on actual size. For pricing information, please see Pricing | Cloud Object Storage.

DEEP ARCHIVE

COS DEEP ARCHIVE is a highly reliable object storage service that offers the lowest storage costs and long-term data retention. This storage class has a minimum storage duration of 180 days. To read data stored in DEEP ARCHIVE, you need to restore it to STANDARD first. For more information, see Overview - DEEP ARCHIVE.
COS supports the following two restoration modes for DEEP ARCHIVE:
Standard: Retrieves an object within 12-24 hours.
Bulk: Retrieves multiple objects within 24-48 hours.
Note:
The QPS of data restoration requests is limited to 100.
Use cases
Use cases that require long-term data retention, such as medical images, view data, and logs.
Note:
DEEP ARCHIVE has minimum storage requirements. If the storage duration is less than 180 days, the bill is calculated as 180 days. Likewise, if the size of a file is smaller than 64 KB, the bill is calculated as 64 KB (if the size of a file is greater than or equal to 64 KB, the bill is calculated based on the actual file size). For more information, see Pricing | Cloud Object Storage.

Storage Class Comparison

Comparison Item
MAZ_STANDARD
MAZ_STANDARD_IA
MAZ_ARCHIVE
MAZ_INTELLIGENT TIERING
INTELLIGENT TIERING
STANDARD
STANDARD_IA
ARCHIVE
DEEP ARCHIVE
Storage class parameter
MAZ_STANDARD
MAZ_STANDARD_IA
MAZ_ARCHIVE
MAZ_INTELLIGENT_TIERING
INTELLIGENT_TIERING
STANDARD
STANDARD_IA
ARCHIVE
DEEP_ARCHIVE
Data durability
99.9999999999%
99.9999999999%
99.9999999999%
99.9999999999%
99.999999999%
99.999999999%
99.999999999%
99.999999999%
99.999999999%
Service availability
99.995%
99.95%
99.50%
It changes with the transition of layers. When in the frequent access tier, service availability can refer to MAZ_STANDARD; when in the infrequent access tier, service availability can refer to MAZ_STANDARD_IA.
It changes with the transition of layers. When in the frequent access tier, service availability can refer to STANDARD; when in the infrequent access tier, service availability can refer to STANDARD_IA.
99.99%
99.90%
99.00%
99.00%
Response
Milliseconds
Milliseconds
Restoration requires advance application. Recovery operations support three modes.
Expedited retrieval mode: Recovery tasks can be completed in 1-5 minutes.
Standard retrieval mode: Recovery tasks complete within 3-5 hours.
Bulk retrieval mode: Recovery tasks complete within 5-12 hours.
Milliseconds
Milliseconds
Milliseconds
Milliseconds
Requires restoration in advance by using one of three restoration modes:
Expedited: Restores an object within 1-5 minutes.
Standard: Restores an object within 3-5 hours.
Bulk: Restores multiple objects within 5-12 hours.
Requires restoration in advance by using one of two restoration modes:
Standard: Restores an object within 12-24 hours.
Bulk: Restores multiple objects within 24-48 hours.
Minimum billable object size
Measured by actual object size
64 KB
64KB
Measured by actual object size
Measured by actual object size
Measured by actual object size
64 KB
64 KB
64 KB
Minimum storage duration
No limit
30 days
90 days
No limit
No limit
No limit
30 days
90 days
180 days
Supported regions
Currently only support Beijing, Shanghai, Guangzhou, Hong Kong (China), Singapore, and Shanghai Finance regions.
Currently only support Beijing, Shanghai, Guangzhou, Hong Kong (China), Singapore, and Shanghai Finance regions.
Currently only support Beijing, Shanghai, Guangzhou, Hong Kong (China), Singapore, and Shanghai Finance regions.
Only Beijing, Shanghai, Guangzhou, and Singapore
Currently, only Beijing, Nanjing, Shanghai, Guangzhou, Chengdu, Chongqing, Singapore, Tokyo, Frankfurt, and Virginia regions are supported.
All regions
All regions

Applicable to public cloud regions (excluding Jakarta) and financial cloud regions
Only Beijing, Nanjing, Shanghai, Guangzhou, Chengdu, Chongqing, Tokyo, and Singapore
Storage fees
High
High
extremely low
Varied by the storage class after intelligent tiering
Varied by the storage class after intelligent tiering
Standard
Low
Very low
Ultra low
Data retrieval fees
None
Rather low. Charged by the actual amount of data read
Relatively high, charge based on actual data volume. Fees vary by recovery mode.
None
None
None
Rather low. Charged by the actual amount of data read
Rather high. Charged by the actual amount of data restored depending on the restoration mode
High. Charged by the actual amount of data restored depending on the restoration mode
Request fees
Standard
Rather high
Standard. Data needs to be restored to MAZ_STANDARD.
Rather high. You will also be charged the INTELLIGENT TIERING object monitoring fees
Rather high. You will also be charged the INTELLIGENT TIERING object monitoring fees
Standard
Rather high
Standard (the data needs to be restored to STANDARD first)
High (the data needs to be restored to STANDARD first). To retrieve DEEP ARCHIVE data, request fees for retrieving data will be incurred
Data processing
Supported
Supported
Supported, but requires recovery first
Supported
Supported
Supported
Supported
Supported (requires data restoration first)
Supported (requires data restoration first)

Storage Class Transition

COS offers various storage classes to indicate how active objects are in COS. You can still change the storage class for objects as needed, or transition objects to a less active storage class such as STANDARD_IA, ARCHIVE, or DEEP ARCHIVE.
Note:
When transitioning objects, ensure that the target storage class is supported for the region where your object resides.
If the object belongs to the archive type (such as ARCHIVE, DEEP ARCHIVE), it is required to restore to the standard type (such as STANDARD) before modifying other storage types. For details, see Restoring Archived Objects.
How each storage class can be transitioned is described as follows:
Storage Class
Available Target Storage Class
Settable Storage Type (Settlement From Hot to Cold, Not Allowed Backward)
MAZ_STANDARD
MAZ_STANDARD_IA, MAZ_ARCHIVE, MAZ_INTELLIGENT_TIERING
MAZ_STANDARD > MAZ_STANDARD_IA> MAZ_INTELLIGENT_TIERING / MAZ_ARCHIVE
MAZ_STANDARD_IA
MAZ_STANDARD, MAZ_ARCHIVE, MAZ_INTELLIGENT_TIERING
MAZ_STANDARD_IA> MAZ_INTELLIGENT_TIERING / MAZ_ARCHIVE
MAZ_ARCHIVE
After restoration, you can change to MAZ_INTELLIGENT_TIERING, MAZ_STANDARD, or MAZ_STANDARD_IA.
None
MAZ_INTELLIGENT_TIERING
MAZ_STANDARD, MAZ_STANDARD_IA, MAZ_ARCHIVE
None
INTELLIGENT TIERING
STANDARD, STANDARD_IA, ARCHIVE, DEEP ARCHIVE
INTELLIGENT TIERING > ARCHIVE > DEEP ARCHIVE
STANDARD
INTELLIGENT TIERING, STANDARD_IA, ARCHIVE, DEEP ARCHIVE
STANDARD > STANDARD_IA > INTELLIGENT TIERING > ARCHIVE > DEEP ARCHIVE
STANDARD_IA
INTELLIGENT TIERING, STANDARD, ARCHIVE, DEEP ARCHIVE
STANDARD_IA > INTELLIGENT TIERING > ARCHIVE > DEEP ARCHIVE
ARCHIVE
INTELLIGENT TIERING, STANDARD, STANDARD_IA, DEEP ARCHIVE (after restoration)
ARCHIVE > DEEP ARCHIVE
DEEP ARCHIVE
INTELLIGENT TIERING, STANDARD, STANDARD_IA, ARCHIVE (after restoration)
None
For detailed directions, see the following documents:
Modifying Storage Classes
Setting Lifecycles


DEEP ARCHIVE Overview

Last updated：2024-03-25 15:28:24

Overview

DEEP ARCHIVE is a storage class provided by COS that enables long-term archiving of massive amounts of data. Its unit storage price is comparable to that of tape storage, making it an ideal low-cost solution for long-term data storage. You can manage data easily at low costs through various interaction means provided by COS, such as API, SDK, tools, and console. This eliminates your needs to maintain complex tape library configuration locally and care about the evolution of underlying storage media.
DEEP ARCHIVE is suitable for data that is accessed infrequently but needs to be retained for a long time. In log cold backup scenarios, enterprises need to back up and store the log data generated every day for tracking and analysis according to applicable laws and regulations. In view data and autonomous driving businesses, enterprises accumulate a high number of media files such as images and videos, which need to be archived and stored persistently after use. DEEP ARCHIVE allows enterprises to store such data in the cloud and restore it only when needed, which reduces the storage costs and simplifies Ops management.
DEEP ARCHIVE is designed for 99.999999999% durability and 99.95% availability. You can also use the versioning and cross-region replication features provided by COS to further guarantee the data security.

How to Use

1. Upload using the console.
1. Log in to the COS console.
2. Click Bucket List and create a bucket in a supported region (such as Beijing).
3. After creating the bucket, click Upload File in File List.
4. Select a local file for upload and select DEEP ARCHIVE for Storage Class in Set Properties. For more information, see Uploading Objects.

5. Upload using APIs. Set x-cos-storage-class to DEEP_ARCHIVE in the PUT Object, POST Object, or Initiate Multipart Upload API for direct upload to the DEEP ARCHIVE storage class.
Note:
Append Object does not support direct upload to DEEP ARCHIVE.
6. Upload using SDKs. Currently, all the SDKs released by COS support direct upload to DEEP ARCHIVE. You need to set the StorageClass parameter to DEEP_ARCHIVE when uploading a file.
7. Upload using a tool. COSBrowser and COSCMD support direct upload to DEEP ARCHIVE. You need to add the header field x-cos-storage-class and set it to DEEP_ARCHIVE when uploading a file.

Use Limits

DEEP ARCHIVE is subject to the following limits. When using it, you need to pay attention to their impact on the storage costs and performance:
Minimum storage unit: The minimum storage unit is 64 KB, and objects smaller than 64 KB will be counted as 64 KB during the calculation of the storage capacity.
Minimum storage period: The minimum storage period is 180 days, and a shorter period will be counted as 180 days for billing.
Note:
The storage period means logical days (a day is 24 hours) and starts from the time when a file is modified.
Retrieval request QPS limit: 100 request/sec.
Available regions: DEEP ARCHIVE is currently supported only in the Beijing, Nanjing, Shanghai, Guangzhou, Chengdu, Chongqing, Tokyo, and Singapore regions and will be available in more regions.
Use limit: Currently, DEEP ARCHIVE is not supported for multi-AZ buckets.
Operation limit: Objects cannot be appended to other objects in the DEEP ARCHIVE storage class.
Retrieval request limit: Only one retrieval request can be executed at a time for an object, and multiple repeated requests will be merged and processed according to the fastest retrieval mode. If the retrieval mode of the (N+1)th request is faster than that of the Nth request, the (N+1)th request will be executed as a new request; otherwise, it will fail.
Data retrieval: Standard retrieval (12–24 hours) and bulk retrieval (24–48 hours) are supported for files in the DEEP ARCHIVE storage class.

Overview - INTELLIGENT TIERING

Last updated：2025-09-16 14:54:01

Overview

INTELLIGENT TIERING is a COS storage class designed to reduce storage costs by automatically moving objects between two storage tiers (frequent access and infrequent access) when access patterns change.
INTELLIGENT TIERING is suitable for data with an unpredictable or variable access pattern. COS monitors the access situation of objects, and storage fees are charged based on the actual storage tier utilized (frequent access tier, infrequent access tier, archive tier, and deep archive tier).Users can change the storage class of objects with uncertain access patterns from STANDARD to INTELLIGENT TIERING as needed to reduce in-cloud storage costs.
Note:
INTELLIGENT TIERING is currently only supported in Beijing, Nanjing, Shanghai, Guangzhou, Chengdu, Chongqing, Singapore, Tokyo, Frankfurt, and Virginia regions.
INTELLIGENT TIERING is a standalone storage class that will incur storage usage fees and object monitoring fees. You can purchase an INTELLIGENT TIERING storage pack to redeem the storage usage fees. For detailed pricing, see Product Pricing.
MAZ buckets do not support enabling configurations for intelligent tiering archive/deep archive tiers.

Advantages

The access patterns of data stored in INTELLIGENT TIERING will be monitored. COS moves objects that have not been accessed for a continuous period to the lower-cost infrequent access tier. If an object in the infrequent access tier is accessed later, it is automatically moved back to the frequent access tier to ensure the data retrieval performance. INTELLIGENT TIERING manages to strike a balance between storage costs and read/write performance and features the following strengths:
Cost-effective: When you use INTELLIGENT TIERING for persistent storage, the longer your data is stored, the more you can save on your storage costs (up to about 20%) compared with the STANDARD storage class. INTELLIGENT TIERING is also involved in the object storage lifecycle, where you can transition your INTELLIGENT TIERING data to ARCHIVE to further lower storage costs.
Stable and durable: INTELLIGENT TIERING provides the same low latency and high throughput as the STANDARD storage class. Besides, it offers durability of up to 99.999999999% (11 9s) by using erasure code-based redundancy, and up to 99.99% availability by using block storage and concurrent reads/writes. INTELLIGENT TIERING has also been integrated with the COS MAZ architecture, featuring high durability (99.9999999999%, that is, 12 9s) and availability (99.995%).
Easy-to-use: To use INTELLIGENT TIERING, all you need to do is specify it as the storage class of your object. This COS storage class is inherently compatible with COS APIs, SDKs, tools, and applications, making it easy for you to manage your in-cloud data as needed.

Supported Storage Tiers

Note:
Files smaller than 64KB will remain in the standard tier and will not descend to the infrequent access tier, archive tier, or deep archive tier.
Users can check which tier an intelligent tiering object is in by using the header x-cos-storage-tier returned by the API HeadObject. Additionally, in the list of objects returned by calling APIs GetBucket and GetBucketVersions, an intelligent tiering object will include the StorageTier field for viewing the object's tier.
Access Tier
x-cos-storage-tier
Frequent Access Tier
FREQUENT
Infrequent Access Tier
INFREQUENT
Archive Tier
ARCHIVE_ACCESS
Deep Archive Tier
DEEP_ARCHIVE_ACCESS

Frequent Access Tier (Enabled by default)

After INT objects are uploaded, they are in the frequent access tier (FREQUENT) by default. While objects are in this tier, storage fees are billed according to the regional standard storage list price.

Infrequent Access Tier (Enabled by default)

When enabling intelligent tiering for a bucket, you need to select the number of days after which to transition to the infrequent access tier, with options including 30, 60, and 90 days. Once set, it cannot be modified.
If an object remains unaccessed for a continuous period of 30 days (or 60 or 90 days), it will transition from the frequent access tier to the infrequent access tier (INFREQUENT). While the object is in this tier, storage fees are billed according to the current region's infrequent storage list price. After being accessed, the object will return to the frequent access tier.

Archive Tier (Optional)

Since objects in the archive tier must be restored before access, enabling the archive tier is optional. Users can enable the archive tier for objects with specific prefixes or tags by adding one or more intelligent tiering archive configuration rules, and set the transition days. The minimum transition period of the archive tier is 91 days, and the maximum is 730 days.
Note:
If the current region does not support archive storage, the console will not allow adding archive configuration rules, and intelligent tiering objects will not descend to the archive tier.
When intelligent tiering objects are in the archive tier, storage fees are charged according to the regional archive storage list price, and there will be no early deletion fees.
Restoring Archive Tier objects
When an object remains unaccessed for N consecutive days, it will transition from the infrequent access tier to the archive tier. Once in the archive tier, data can only be read after the object is restored via PostObjectRestore.
Unlike regular archive types, restoring an object from an INT archive tier will not create a standard-type replica. Instead, the object itself will directly return to the frequent access tier.
Similar to the regular archive type, INT archive tier objects support three retrieval modes: expedited, standard, and bulk.
Billing
INT archive retrieval fee:
Standard retrieval and bulk retrieval are free of charge.
Expedited retrieval billing follows the same pricing as intra-regional archive storage expedited retrieval fee.
INT archive retrieval request fee:
Standard retrieval requests and bulk retrieval requests are free of charge.
Expedited retrieval request billing follows the same pricing as intra-regional archive storage expedited retrieval request fee. Since archive storage expedited retrieval requests are currently free of charge, INT archive expedited retrieval requests also remain free of charge to maintain consistency.

Deep Archive Tier (Optional)

Similarly, deep archive tier objects need to be restored before access. Therefore, enabling the deep archive tier is optional. Users can activate the deep archive tier for objects with specific prefixes or tags by adding one or more intelligent tiering deep archive configuration rules. It is possible to configure both the archive tier and deep archive tier transition in the same rule. It is important to note that the transition days for the deep archive tier must be at least 180 days and a maximum of 730 days, and must be greater than the transition days for the archive tier.
Note:
If the current region does not support deep archive storage, the console will not allow adding deep archive configuration rules, and intelligent tiering objects will not descend to the deep archive tier.
When INT objects are in the deep archive tier, storage fees are charged according to the regional deep archive storage list price, and there will be no early deletion fees.
Restoring objects from the Deep Archive Tier
When an object remains unaccessed for M days, it will transition from the infrequent access tier/archive tier to the deep archive tier. Once in the deep archive tier, data can be read only after the object is restored through PostObjectRestore.
Unlike regular deep archive types, restoring an object from the INT Deep Archive Tier will not produce a standard-type replica; instead, the object itself will directly return to the frequent access tier.
Similar to regular deep archive types, INT deep archive tier objects support two retrieval modes: standard and bulk, and there will be neither retrieval fees nor retrieval request fees.

Enabling Intelligent Tiering, Archive Tier and Deep Archive Tier Configuration

Rule Description

Once a bucket is enabled with intelligent tiering configuration, the uploaded intelligent tiering objects by default only switch between the frequent access tier and the infrequent access tier. After archive and deep archive tier rules are added in the intelligent tiering configuration, switching to the archive and deep archive tiers can be enabled. Each bucket can support up to 1000 archive and deep archive tier configuration rules, with the elements included in the rules as follows. For more details on the rules, refer to the API documentation PUT Bucket IntelligentTiering.

Rule Name (Id)

Used to uniquely identify archive and deep archive rules.

Status

The archive tier and deep archive tier configuration supports enabling (Enabled) or disabling (Disabled). Even if archive/deep archive tier rules are configured, they will not be actually converted to the archive or deep archive tier when disabled.

Application Scope (Filter)

The effective range of the rules. It supports prefix filtering and tag filtering. The number of prefixes cannot exceed one, and the number of tags cannot exceed ten.

Hierarchy Settings

Supports setting the transition time for the archive tier and deep archive tier within a single rule. After the corresponding rules are set, INT objects will start the transition to the archive and deep archive tiers. Otherwise, they will only switch between the frequent access tier and the infrequent access tier.
Archive Tier (ARCHIVE_ACCESS): Supports the setting between 91 to 730 days.
Deep Archive Tier (DEEP_ARCHIVE_ACCESS): Supports the setting between 180 to 730 days.
Note:
If both archive tier and deep archive tier conversion rules are configured in one rule, the conversion days for the deep archive tier must be greater than those for the archive tier.

Storage Tier Transition Order

Users can enable archive tier transition and deep archive tier transition separately or enable both at the same time. The transition will proceed sequentially based on the configured days for the infrequent access tier, archive tier, and deep archive tier.
Configuration Samples
Configuration Details
Effective Result
Sample 1
30 days for the infrequent access tier, 100 days for the archive tier, and 190 days for the deep archive tier
After 30 consecutive days of no access, transition from the standard access tier to the infrequent access tier will occur.
After 100 consecutive days of no access, transition from the infrequent access tier to the archive tier will occur.
After 190 consecutive days of no access, transition from the archive tier to the deep archive tier will occur.
Example 2
30 days for the infrequent access tier, and 190 days for the deep archive tier
After 30 consecutive days of no access, transition from the standard access tier to the infrequent access tier will occur.
After 190 consecutive days of no access, transition from the infrequent access tier to the deep archive tier will occur.
Example 3
60 days for the infrequent access tier and 91 days for the archive tier
After 60 consecutive days of no access, transition from the standard access tier to the infrequent access tier will occur.
After 91 consecutive days of no access, transitions from the infrequent access tier to the archive tier will occur.

Restoring INT Objects from the Archive and Deep Archive Tiers

When an object remains unaccessed for many consecutive days, it will be transitioned from the infrequent access tier to the archive/deep archive tier. Once it enters the archive or deep archive tier, it needs to be restored using PostObjectRestore before the data can be read.
Unlike the restoration of regular archive types and deep archive types of objects, the restoration of INT archive and deep archive tier objects does not create a standard replica; instead, the object itself will directly return to the frequent access tier. Therefore, when initiating a retrieval request for INT archive/deep archive tier objects, there is no need to specify the number of days for retrieval.
Similar to the regular archive types, INT archive tier objects support expedited, standard, and bulk retrieval modes; like the regular deep archive types, INT deep archive tier objects support standard and bulk retrieval modes.
In terms of charges, apart from the archive tier expedited retrieval fee (priced the same as the regional archive storage expedited retrieval fee), no other retrieval fees, retrieval request fees, or rewarming replica storage fees are charged.

Archive Storage Type
Intelligent Tiering Object - Archive Tier
Deep Archive Storage Type
Intelligent Tiering Object - Deep Archive Tier
Storage Fee
Storage fees are charged according to the current regional archive storage type
Storage fees are charged according to the current regional deep archive storage type
Whether retrieval generates a replica
Yes, it is necessary to specify the retention period, and replica storage fees are incurred according to standard storage pricing.
No, there will be no replica storage fees.
Yes, it is necessary to specify the retention period, and replica storage fees are incurred according to standard storage pricing.
No, there will be no replica storage fees.
Retrieval Mode
Expedited mode
Standard mode
Bulk mode
Expedited mode
Standard mode
Bulk mode
Standard mode
Bulk mode
Standard mode
Bulk mode
Retrieval Fee
(In USD/GB)
Expedited
ARCHIVE data retrieval fees (expedited)
Expedited Retrieval Fee for Intelligent Tiering Archive Tier (Same Pricing as Archiving Storage)
\
\
Standard
ARCHIVE data retrieval fees (standard)
Free
DEEP ARCHIVE data retrieval fees (standard)
Free
Bulk
ARCHIVE data retrieval fees (bulk)
Free
DEEP ARCHIVE data retrieval fees (bulk)
Free
Retrieval Request Fees
(In USD/10,000 Calls)
not billed
not billed
Deep Archive Standard Retrieval Request Fee Deep Archive Bulk Retrieval Request Fee
Free
Read/Write Request Fee
Consistent with the standard storage fee after the restoration
Intelligent Tiering Read and Write Request Fee
Deep Archive Read and Write Request Fee
Intelligent Tiering Read and Write Request Fee
During the restoration process, you can use HeadObject to check the restoration status of INT archive tier/deep archive tier objects.
When in the process of restoration, the response header of HEAD Object will include x-cos-restore and x-cos-restore-status. For example, x-cos-restore:ongoing-request="true", cos-restore-status:tier="bulk"; request-date="Mon, 18 Nov 2019 09:34:50 GMT".
After the restoration is complete, INT objects directly return to the frequent access tier, and the response header `x-cos-storage-tier` of HEAD Object is `FREQUENT`.

How to Use

To store your object in the COS INTELLIGENT TIERING storage class, enable INTELLIGENT TIERING for the bucket, and then specify the storage class of your object as INTELLIGENT TIERING.

Using COS Console

Setting INTELLIGENT TIERING upon object uploads

1. Log in to the COS Console.
2. In the left menu bar, click bucket list to enter bucket list page.
3. Locate the bucket that needs intelligent tiered storage, click its name, and enter the bucket management page.
4. Click Basic Configuration > Intelligent Tiering, locate the intelligent tiering storage configuration item, click Edit to toggle on the current status, and configure according to the configuration item description.
Note:
Once intelligent tiering storage is configured, this configuration cannot be closed or suspended.
Low-frequency layer conversion days: This parameter is used to specify the time when converting to the infrequent access tier. Selectable number of days are 30, 60, and 90. For example, when set to 30 days, the system will move objects from the frequent access tier to the infrequent access tier after 30 consecutive days of no access.
5. After confirming that the configuration information is correct, click Save, complete the secondary confirmation according to the prompts, and intelligent tiering can be enabled.
6. Once enabled, intelligent tiering storage supports configuring archiving and deep archiving rules. Intelligent tiered objects within the specified range can be converted to the archive tier and deep archive tier. If you do not need to configure intelligent tiered archiving and deep archiving rules, skip to Procedure 10.

7. Click Add Rule. Configuration items are described below:

Rule name: Enter your rule name. Once created, it cannot be modified.
Application scope: This intelligent tiering archive and deep archiving rules can be applied to the entire bucket or objects in a specified range. Currently supported ranges are:
Object prefix: Specify objects with the same file prefix to execute intelligent tiering archive and deep archiving rules, such as prefix/.
Object tag: Specify objects with the same tag to execute intelligent tiering archive and deep archiving rules. A maximum of 10 tags can be specified. Case-sensitive.
Note:
You can specify both object prefix and object tag. The relationship between object prefix and object tag, or between object tags, is "AND" (all conditions must be satisfied at the same time). For example, if you set the object prefix to doc and the object tag key-value pair to group = IT in intelligent tiering archive and deep archiving rules, the specified objects will be all objects in the current bucket with the object key prefix doc and the object tag group = IT.
Archive rules: Supports configuration of the number of days for intelligent tiered objects to be switched to archive and deep archive. Configure at least one.
8. After confirming the information, click Confirm, and you will see the intelligent tiering archive and deep archiving rules.

Note:
The effective result of the rules follows the configured conversion days for infrequent, archive, and deep archive tiers, settling sequentially.
For example, a user configures three rules: set the low-frequency tier conversion days to 30, set the archive tier conversion days to 100, and set the deep archive tier conversion days to 190. The effective result is: if an intelligent tiering object in the specified range has no access for 30 consecutive days, it converts from the standard tier to the low-frequency tier; if no access for 100 consecutive days, it converts from the low-frequency tier to the archive tier; if no access for 190 consecutive days, it converts from the archive tier to the deep archive tier.
9. To stop the intelligent tiering archive and deep archiving rules, click Edit, change the corresponding rule status to Disabled, or just delete the lifecycle rule.
10. 
After enabling the intelligent tiering storage configuration, click File List in the left menu bar.

11. On the File List page, click Upload File.
12. In the pop-up window, select upload file, click Parameter Configuration to set object properties, and choose Intelligent Tiering Storage in the storage type configuration.
13. Click Upload to upload objects as Intelligent Tiered Storage Type. COS will automatically conduct storage tier conversion. For other configuration items about upload operation, please refer to Upload Object.

Moving in-cloud objects to INTELLIGENT TIERING

You can perform the following steps to move in-cloud objects to the INTELLIGENT TIERING storage class:
1. On the bucket configuration page, create a lifecycle rule. For detailed directions, see Setting Lifecycle.
2. Specify the application range and transition objects to INTELLIGENT TIERING.

Using the REST API

You can use the following APIs to configure INTELLIGENT TIERING:
1. Use RESTful APIs to enable INTELLIGENT TIERING for your bucket. For more information, see the following API documentation:
To enable intelligent tiering and set the transition days for the infrequent access tier:
PUT Bucket IntelligentTiering (id is not default)
GET Bucket IntelligentTiering (id is not default)
To configure or delete archive and deep archive tier rules:
PUT Bucket IntelligentTiering (id is not default)
GET Bucket IntelligentTiering (id is not default)
DELETE Bucket IntelligentTiering (id is not default)
2. After enabling INTELLIGENT TIERING for the bucket, use the following APIs to store the object in the INTELLIGENT TIERING storage class:
PUT Object
PUT Object - Copy
POST Object
Initiate Multipart Upload
3. You can use the following APIs to query the storage class or storage tier of an object:
GET Object
HEAD Object
4. You can use the following RESTful APIs to delete objects in the INTELLIGENT TIERING storage class:
DELETE Object
DELETE Multiple Objects
PUT Bucket lifecycle
DELETE Bucket IntelligentTiering
5. To retrieve objects from the intelligent tiering archive tier and deep archive tier, please refer to the following API: PostObjectRestore.

Using the SDK

Currently, all COS SDKs support INTELLIGENT TIERING. To use this storage class, set StorageClass to INTELLIGENT_TIERING or MAZ_INTELLIGENT_TIERING when uploading an object. For more information, see Uploading Object.

Use Limits

INTELLIGENT TIERING has the following limits:
Setting: The setting cannot be changed once configured. To change it, contact us. The number of consecutive no-access days for moving objects to the infrequent access tier can be set to 30, 60, or 90.
Initial storage tier: A new object is stored in the frequent access tier by default, and will only be moved to the infrequent access tier if it hasn’t been accessed for a certain period.
Minimum storage size: An object smaller than 64 KB can only be stored in the frequent access tier and cannot be moved between the frequent and infrequent access tiers. Objects are billed based on their actual sizes.
Operation: Uploading objects to INTELLIGENT TIERING using the APPEND Object API is not supported.
Lifecycle: Objects in the INTELLIGENT TIERING storage class can only be transitioned to ARCHIVE or DEEP ARCHIVE. If a STANDARD object is transitioned to INTELLIGENT TIERING, it will be stored in the frequent access tier. If a STANDARD_IA object is transitioned to INTELLIGENT TIERING, it will be stored in the infrequent access tier.
Bucket replication: During bucket replication, if INTELLIGENT TIERING is not enabled for the destination bucket, the object cannot be copied to the INTELLIGENT TIERING storage class.

FAQs

How is INTELLIGENT TIERING billed?

INTELLIGENT TIERING fees include INTELLIGENT TIERING storage usage fees and INTELLIGENT TIERING object monitoring fees.
INTELLIGENT TIERING storage usage fees are charged differently depending on the storage tier of objects.
When objects are stored in the frequent access tier, STANDARD storage usage fees are charged.
When objects are stored in the infrequent access tier, STANDARD_IA storage usage fees are charged.
Note:
STANDARD and STANDARD_IA storage usage fees vary by region. For details about their pricing, see Product Pricing.
Request fees and traffic fees are also incurred during object upload and download. For the billing examples, see Traffic Fee Billing Example and Request Fee Billing Example.
INTELLIGENT TIERING object monitoring fees are charged based on the number of objects stored (excluding files smaller than 64 KB). For detailed pricing, see Pricing | Cloud Object Storage.
Example
Suppose an organization has 100,000 objects (all above 64 KB, 1 TB in total), and the data is stored in the INTELLIGENT TIERING storage class in Beijing region and transitioned to the infrequent access tier after 30 days. If 20% of the objects (i.e., 20,000 objects) are transitioned to the infrequent access tier every 30 days, the object monitoring fees and storage usage fees for every 30 days will be as follows:
Note: In the table below, the object monitoring fees for Beijing region are 0.25 USD/10,000 objects per 30 days, that is, 0.00833333 (0.25 / 30) USD/10,000 objects per day.
Storage Days
30-day Object Monitoring Fees (USD)
30-day INTELLIGENT TIERING Storage Usage Fees (USD)
30-day STANDARD Storage Usage Fees (USD)
30 * 1
0.25/10,000 objects * 100,000
1024 * 0.024 / 30 * 30 = 24.58
1024 * 0.024 / 30 * 30 = 24.58
30 * 2
0.25/10,000 objects * 100,000
819.2 * 0.024 / 30 * 30 + 204.8 * 0.08 / 30 * 30 = 23.35
1024 * 0.024 / 30 * 30 = 24.58
30 * 3
0.25/10,000 objects * 100,000
655.36 * 0.024 / 30 * 30 + 368.64 * 0.08 / 30 * 30 = 22.36
1024 * 0.024 / 30 * 30 = 24.58
30 * 4
0.25/10,000 objects * 100,000
524.288 * 0.024 / 30 * 30 + 499.712 * 0.08 / 30 * 30 = 21.58
1024 * 0.024 / 30 * 30 = 24.58
30 * 5
0.25/10,000 objects * 100,000
419.4304 * 0.024 / 30 * 30 + 604.5696 * 0.08 / 30 * 30 = 20.95
1024 * 0.024 / 30 * 30 = 24.58
30 * 6
0.25/10,000 objects * 100,000
335.54432 * 0.024 / 30 * 30 + 688.45568 * 0.08 / 30 * 30 = 20.45
1024 * 0.024 / 30 * 30 = 24.58
As you can see, using INTELLIGENT TIERING reduces storage costs over time. Only a small amount of monitoring cost is required every 30 days.

What types of objects is INTELLIGENT TIERING suitable for?

INTELLIGENT TIERING is suitable for large objects (such as audio and video, logs, etc.) whose access patterns change. Larger average object sizes mean you pay less for monitoring per GB objects. If your business has relatively fixed data access patterns, you can set lifecycle configuration to specify the time to transition objects to STANDARD_IA without the need to use INTELLIGENT TIERING.

How can I store objects in INTELLIGENT TIERING?

You can store objects in INTELLIGENT TIERING as follows:
Incremental objects: You just need to specify the storage class as INTELLIGENT TIERING when uploading the objects.
Existing objects: You can modify the storage class to INTELLIGENT TIERING using the COPY API. You can also use the lifecycle feature to transition STANDARD or STANDARD_IA objects to INTELLIGENT TIERING.
Note:
In INTELLIGENT TIERING, objects smaller than 64 KB will always be stored in STANDARD. Therefore, for objects smaller than 64 KB, we recommend you upload them to the STANDARD, STANDARD_IA, ARCHIVE, or DEEP ARCHIVE storage class as needed to reduce costs.

How do I disable INTELLIGENT TIERING configuration?

INTELLIGENT TIERING configuration can't be disabled once enabled. If you don't need to store your objects in INTELLIGENT TIERING, you just need to specify the storage class as a non-INTELLIGENT TIERING class such as STANDARD, STANDARD_IA, ARCHIVE, or DEEP ARCHIVE when uploading the objects.
The rules for intelligent tiering archive tier/deep archive tier support deletion. You can delete the corresponding rules to prevent new INT objects from descending into the archive tier/deep archive tier.

How are charges applied for archiving in the intelligent tiering and deep archive tier?

Storage fee: Charged according to the regional archive and deep archive storage fees.
Retrieval fee and retrieval request fee: Apart from the expedited retrieval fees for the archive tier, no other retrieval fees nor retrieval request fees are charged.
Rewarming replica fee: Not charged.

Direct Upload to ARCHIVE

Last updated：2024-01-06 15:06:14

Overview

COS supports direct upload to the ARCHIVE storage class by using the console, APIs, SDKs, or COSCMD. Alternatively, you can use the lifecycle feature of COS to transition your objects to the ARCHIVE or DEEP ARCHIVE storage class. For more information, see Storage Class Overview.

Supported Methods

Uploading via the console In the COS console, click Upload Files and select the object to be uploaded. In the Set Properties step, set Storage Class to ARCHIVE or DEEP ARCHIVE. For more information, see Uploading Object.

Uploading via APIs You can set x-cos-storage-class to ARCHIVE or DEEP_ARCHIVE in the PUT Object, POST Object, or Initiate Multipart Upload API to implement direct upload to ARCHIVE.
Note:
The Append Object API does not support direct upload to ARCHIVE.
Uploading via SDKs Currently, all COS SDKs support direct upload to ARCHIVE. You can set the StorageClass parameter to ARCHIVE or DEEP_ARCHIVE during the upload.
Uploading via COSCMD You can add the header field x-cos-storage-class and set it to ARCHIVE or DEEP_ARCHIVE during the upload.

Restoring and downloading archived data

Unlike STANDARD/STANDARD_IA, objects in ARCHIVE or DEEP ARCHIVE can only be downloaded after being restored. The following three restoration modes are provided:
Expedited: Restores an object within 1–5 minutes (fastest).
Standard: Restores an object within 3–5 hours.
Bulk: Restores multiple objects within 5–12 hours (most cost-effective).
Note:
Objects in DEEP ARCHIVE cannot be restored in the expedited mode. An object can be restored within 12–24 hours in standard mode or 24–48 hours in bulk mode.
Note that the console, APIs, SDKs, and COSCMD all support the restoration and download of archived objects.

Restrictions

Objects in ARCHIVE or DEEP ARCHIVE need to be restored before being downloaded.
To copy an object in ARCHIVE or DEEP ARCHIVE, you need to restore it first.
Objects in ARCHIVE and DEEP ARCHIVE do not support cross-region replication.
Objects in ARCHIVE and DEEP ARCHIVE cannot be changed to a more frequently accessed storage class, such as STANDARD and STANDARD_IA.

Modifying Storage Class

Last updated：2025-11-10 17:33:34

Overview

You can modify the storage class of your object uploaded to COS through the console at any time to meet your business needs in different scenarios. COS offers a range of storage classes, including MAZ_STANDARD / STANDARD, MAZ_STANDARD_IA / STANDARD_IA, MAZ_INTELLIGENT TIERING / INTELLIGENT TIERING, ARCHIVE / ARCHIVE, and DEEP ARCHIVE. For more information, see Overview. The following section will guide you through how to modify the storage class of an object.
Note:
If your objects are in Multi-AZ storage types such as MAZ_STANDARD or MAZ_STANDARD_IA, changing the storage type to non-Multi-AZ types (such as STANDARD_IA or ARCHIVE) is not currently supported.
If an object belongs to archive type (such as ARCHIVE or DEEP_ARCHIVE), it is required to restore to standard type (such as STANDARD) before changing to other storage types. For more information, see Restoring Archived Objects.
Modifying the storage class of objects larger than 5 GB is not supported.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Locate the bucket where the object resides and click the bucket name.
4. Click File List on the left sidebar.
5. Find the target object and click More > Modify Storage Class in the Operation column on the right.
6. To batch modify the storage class of multiple objects, select multiple objects and click More Actions > Modify Storage Class at the top.
7. Select the target storage class and click OK in the pop-up window.

Uploading Object

Simple Upload

Last updated：2025-09-23 15:22:11

Simple upload refers to users uploading objects using the PUT Object interface. This method is suitable for uploading individual objects not exceeding 5GB in a single request.
To upload individual objects larger than 5GB, you can use the following methods:
Using the console upload: The console supports uploading individual objects up to 512GB in size. For more information, please refer to the Upload Object Console Operation Guide.
Using API/SDK for multipart upload: This method supports uploading individual objects up to 48.82TB (i.e., 50,000GB). Please refer to Multipart Upload.
Using the COSCMD tool, you can upload individual objects up to 40TB in size. For more information, please refer to COSCMD Tool.
Note
When initiating an upload request, if you need to upload to a specific folder or path, you can use / to achieve this. For example, to upload picture.png to the doc folder, set the object key as doc/picture.png.

Scenarios

The simple upload method is suitable for scenarios where the object to be uploaded does not exceed 5GB.
In a high bandwidth or weak network environment, if your object is relatively large, for example, over 100 MB, we recommend you use Multipart Upload, even if the object is smaller than 5 GB. This is because multipart upload supports uploading multiple parts concurrently, which gives full play to resources in high bandwidth environments. On the other hand, in weak network environments, the uploaded parts will not be affected by the failed ones. Therefore, failed parts can be re-uploaded with a simple retry, improving the overall upload success rate. For more information about uploading objects with a client in a weak network environment, see Multipart Upload Resumption in a Weak Network Environment.

How to Use

Using the Console

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the target bucket to enter the File List page.
4. In File List, click Upload Files.
5. In the pop-up window, click Select Files or Select Folders and select one or multiple local files (or folders) as needed.
Note:
Since some browsers do not support uploading multiple files, we recommend you use popular browsers such as Internet Explorer 10 or later, Firefox, or Chrome.
6. (Optional) Click Configure Parameters.



7. (Optional) set the object attributes in the Upload Files window.



The configuration instructions are as follows:
Storage Class: Select the storage class for your object as needed. This field is set to STANDARD by default. For more information, see Overview.
Note:
If your bucket has MAZ configuration enabled, you can only select an MAZ storage class, such as MAZ_STANDARD. If it also has INTELLIGENT TIERING configuration enabled, you can also select MAZ_INTELLIGENT TIERING.
Access Permissions: Select the access permission for your object as needed. This field is set to Inherit by default (inheriting permissions of the bucket). For more information, please see Basic Concepts of Access Control.
Server-side encryption: When you upload objects, configure the server encryption attribute for the objects. Tencent Cloud COS will add a data encryption protection policy for the uploaded objects, helping you automatically encrypt before data is written and automatically decrypt when you access the data. Currently, SSE-COS and SSE-KMS are supported. For an introduction to server-side encryption and applicable regions, see Server-side Encryption Overview.
Object tag: You can add tags to an object to be uploaded. The object tag is composed of a tag key, (=), and a tag value, such as group = IT. Each object tag is a key-value pair. For more information, see Object Tag Overview.
Metadata: Object metadata, or HTTP header, is a string sent by the server over HTTP before it sends HTML data to the browser. By modifying HTTP headers, you can modify how the webpage responds as well as certain configurations, such as caching time. Modifying an object's HTTP headers does not modify the object itself. For more information, please see Custom Headers.
8. Click Upload.
9. You can check the upload progress in Task Completed in the top-right corner of the page. Once the upload is complete, the uploaded object will appear in File List.
Note:
The task progress in the figure indicates the number of tasks established by the current operation. For example, if you have executed an upload task (this upload task is to upload 10 files) and all have been uploaded successfully, then the task progress will display "Task completed (10 successful)".

Using the REST API

Use REST APIs to initiate a simple upload request. For more information, see PUT Object.

Using SDKs

Directly call the simple upload method in the SDK. For more information, see the SDK documentation for the corresponding programming language below:
Android SDK
C SDK
C++ SDK
.NET SDK
Go SDK
iOS SDK
Java SDK
JavaScript SDK
Node.js SDK
PHP SDK
Python SDK
Mini Program SDK

Multipart Upload

Last updated：2024-03-25 15:28:24

Overview

Multipart upload can split an object into multiple parts and upload them to COS. When uploaded, the parts are numbered consecutively. You can upload each part separately or upload them in any order. COS will merge them into an object based on their numbers. If the upload of any part fails, the failed part can be uploaded again without affecting other parts and the content as a whole. In a poor network environment, multipart upload is recommended for objects larger than 20 MB. In a high-bandwidth environment, multipart upload is recommended for objects larger than 100 MB.
Multipart upload can split an object into up to 10,000 parts of 1 MB to 5 GB in size each. The last part can be smaller than 1 MB.
Note:
Simple upload only supports uploading files with a maximum size of 5 GB, while multipart upload supports larger files.

Use Cases

Multipart upload is suitable for uploading large objects in poor or high-bandwidth network environments.
Multipart upload has the following strengths:
In a poor network environment, smaller part size minimizes the impact of restarting a failed upload due to network errors.
In a high-bandwidth environment, multipart upload can maximize the use of your available bandwidth by uploading object parts in parallel. Uploading parts out of order does not affect the final merged object.
With multipart upload, you can pause and resume the upload of a single large object at any time. All incomplete multipart uploads can be resumed unless aborted.
Multipart upload can also be used to begin an upload before you know the final object size. You can initiate an upload and then merge the parts to get the full object size.

How to Use

Using RESTful API

You can use a RESTful API directly to initiate a multipart upload request. For more information, see the following API documents:
Initiate Multipart Upload
Upload Part
Complete Multipart Upload
Abort Multipart Upload

Using SDK

You can directly call the multipart upload method in the SDK. For more information, see the SDK documentation for the corresponding programming language below:
Android SDK
C SDK
C++ SDK
.NET SDK
Go SDK
iOS SDK
Java SDK
JavaScript SDK
Node.js SDK
PHP SDK
Python SDK
Mini Program SDK

Upload via Pre-Signed URL

Last updated：2025-08-27 11:38:11

Use Cases

All buckets and objects are private by default. If you want any third party to be able to upload an object to your bucket, but you don't want them to use CAM accounts or temporary keys, signatures can be provided by pre-signed URLs for temporary upload operations. Anyone who receives a valid pre-signed URL can upload an object.
When creating a pre-signed URL, you can include an object key in your signature so that the object can only be uploaded to the specified object key. Besides, the validity period of pre-signed URLs can be provided in SDKs to ensure that expired URLs will not be used by any unauthorized party.
Note:
Accessing a CDN domain needs to follow the authentication process of CDN, in which case COS signatures cannot be used. Therefore, pre-signed URLs do not support the use of CDN domains.
Use a permanent key to generate a pre-signed URL, noting that the permission scope is limited to upload or download operations, and the validity period of the generated signature is set to the shortest duration required to complete the upload or download operation.
Once the specified validity period of the pre-signed URL expires, the request will be interrupted; a new signature must be applied for, and the failed request needs to be retried.
Pre-signed URLs only support pre-signing for a single object and do not support multiple objects.
The pre-signed URL upload method only supports simple uploads, with a maximum file size of 5GB. Chunked uploads (resume uploads) are not supported.

Directions

Using SDKs

You can call the pre-signed URL method in the SDK directly. For more information, please see the SDK documentation for the corresponding programming language below:
Android SDK
C SDK
C++ SDK
.NET SDK
Flutter SDK
Go SDK
iOS SDK
Java SDK
JavaScript SDK
Node.js SDK
PHP SDK
Python SDK
React Native SDK
Mini Program SDK

Downloading Object

Simple Download

Last updated：2025-10-13 16:29:32

Scenarios

You can directly initiate a request to download objects in COS. The following features are supported:
Download a complete object: Download the complete object data by initiating a GET request.
Download a part of an object: Use the Range request header in a GET request to retrieve a specific range of bytes of an object. Retrieving multiple ranges is not supported.
The object's metadata will be returned as HTTP response headers along with the object content. GET requests support overriding some metadata values using URL parameters, such as the Content-Disposition response value. Modifiable response headers include:
Content-Type
Content-Language
Expires
Cache-Control
Content-Disposition
Content-Encoding

How to Use

Using COS console

Downloading a single object

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Locate the bucket where the object resides and click the bucket name.
4. Click File List on the left.
5. Find the target object and click the corresponding Download button.
Click Download in the Operation column on the right of the object.
Select the object and click More Actions > Download at the top.
Click Details in the Operation column on the right of the object to enter the file details page, and then click Download Object.
You can also click Copy Temporary Link, paste the link into a browser, and press Enter.
Note:
If the bucket where the object is stored is private read/write, a signature will be automatically calculated and added at the end of the address copied here. For more information on how to generate a signature, see Request Signature.
The temporary link with a signature is valid for 1 hour from the moment you click to view the Details. You can refresh the validity period of the signature by clicking Refresh.
By default, if the downloaded object can be directly opened in the browser, then accessing the temporary URL will directly preview the object rather than downloading it.

Downloading objects or folders in batches

Note:
You can only download individual objects in the COS console. To download multiple objects or folders in batches, we recommend you install the COSBrowser client. Here is how to download objects or folders in batches in the console in conjunction with the COSBrowser client.
1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Locate the bucket where the object resides and click the bucket name.
4. Click File List on the left.
5. Select multiple objects and click More Actions > Download at the top.
6. Follow the on-screen prompts to install or launch the COSBrowser client and log in.
7. After opening COSBrowser, select the file storage location. Selected files will automatically proceed to the download queue and start the download. You can click the upper right Transmission icon to view and download progress.

Using the REST API

Use the REST API to initiate an object download request. For more information, see GET Object.

Using SDKs

Directly call the object download method in the SDK. For more information, see the SDK documentation for the corresponding programming language below:
Android SDK
C SDK
C++ SDK
.NET SDK
Go SDK
iOS SDK
Java SDK
JavaScript SDK
Node.js SDK
PHP SDK
Python SDK
Mini Program SDK


Download via Pre-Signed URL

Last updated：2025-08-29 10:47:09

Overview

All buckets and objects are private by default. If you want any third party to be able to download an object without using CAM account or temporary keys, provide the third parties with signatures via pre-signed URLs for download operations. Anyone who receives a valid pre-signed URL can download an object.
When creating a pre-signed URL, you can include object keys in your signature to specify the objects allowed for download. Besides, the validity period of pre-signed URLs can be provided in SDKs to ensure that expired URLs will not be used by any unauthorized party.
Note:
Accessing a CDN domain needs to follow the authentication process of CDN, in which case COS signatures cannot be used. Therefore, pre-signed URLs do not support the use of CDN domains.
You are advised to use a temporary key to generate pre-signed URLs for the security of your requests such as uploads and downloads. When you apply for a temporary key, follow the Principle of Least Privilege to avoid leaking resources besides your buckets and objects.
If you need to use a permanent key to generate a pre-signed URL, you are advised to limit the permission of the permanent key only to uploads and downloads to avoid risks. In addition, the validity period of the generated URL must be set to the shortest period required to complete the current upload or download operation.
Once the validity period of the specified pre-signed URL expires, the request will be interrupted; after applying for a new signature, the failed request needs to be retried, and resuming from a breakpoint is not supported.
The pre-signed URL only supports pre-signing for a single object and does not support multiple objects.

Directions

Using SDKs

Call the pre-signed URL method in the SDK. For more information, see the SDK documentation for the corresponding programming language below:
Android SDK
C SDK
C++ SDK
.NET SDK
Flutter SDK
Go SDK
iOS SDK
Java SDK
JavaScript SDK
Node.js SDK
PHP SDK
Python SDK
React Native SDK
Mini Program SDK

Deleting Objects

Deleting Objects

Last updated：2025-09-26 11:18:44

Overview

Tencent Cloud Object Storage (COS) supports deleting one or more objects uploaded to a bucket. You can delete objects in batches through the console, API, or SDK. By default, an empty response is returned if the task deletion completes successfully. If an error occurs, an error message is returned.
Note:
A maximum of 1000 objects can be deleted per request. If needed, split the list and send the request separately.
If you select delete "folder", it will trigger simultaneous deletion of all objects under the "folder". With Versioning Not Enabled, data deletion cannot be recovered or accessed.
If versioning is enabled for your current bucket (including enabled and paused states), performing deletion operation does not actually delete the object but adds a deletion marker to the current version object. At the same time, your storage usage will not reduce. To truly delete an object, you need to specify a version ID. For more information, please refer to Version Control Overview.
If you need to delete large amounts of data or nested folders, it may take a long time to complete the deletion. For such scenarios, it is recommended to set lifecycle for deletion.

Applicable Scenarios

Tencent Cloud Object Storage (COS) supports directly deleting one or multiple objects. To delete an object, you only need to provide its object key and call an API request to delete it.

Usage

Using the COS Console

Deleting an Object in a Bucket with Versioning Not Enabled

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to go to the bucket list page.
3. Find the bucket where the object is located and click the bucket name to enter the bucket management page.
4. In the left sidebar, select File List to go to the file list page.
5. Select one or more objects you want to delete, click More Operations > Delete at the top.
6. In the pop-up prompt window, click Delete to delete the object.

Deleting an Object in a Bucket with Versioning Enabled/Suspended

Note:
If the bucket has used the version control feature, you must specify a Version ID to actually delete an object. If you delete an object without specifying a Version ID, the object will not be actually deleted. Instead, a deletion marker will be added to the current version. For more information, please refer to Version Control Overview.
1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to go to the bucket list page.
3. Find the bucket where the object is located and click the bucket name to enter the bucket management page.
4. In the left sidebar, select File List to go to the file list page.
5. Turn on the "Show Noncurrent Versions" option at the top, then select one or more object versions to delete, and click More Operations > Delete at the top.
6. In the pop-up prompt window, click Delete to actually delete the object.

Using the REST API

You can use the REST API to initiate deletion requests for one or more objects. See the DELETE Object and DELETE Multiple Objects API documents.

Using SDKs

Directly call the single object deleting method in the SDK. For more information, see the SDK documentation for the corresponding programming language below:
Android SDK
C SDK
C++ SDK
.NET SDK
Go SDK
iOS SDK
Java SDK
JavaScript SDK
Node.js SDK
PHP SDK
Python SDK
Mini Program SDK

Deleting Incomplete Multipart Uploads

Last updated：2025-11-19 15:11:10

Overview

If you try to delete the specified bucket but the system prompts that "Deletion failed. Please delete the valid data in the bucket first", you can enter Incomplete Multipart Upload to view the files that have not been completely uploaded and delete them. The bucket can be deleted only after you confirm that all completely and partially uploaded files have been deleted from the bucket.
Note:
During the object upload progress, the files that are paused or canceled will be displayed in Incomplete Multipart Upload. Files can be viewed in File List only after they are completely uploaded.
Like objects, incomplete multipart uploads consume your storage space and incur storage usage costs.

Directions

Deleting incomplete multipart uploads manually

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Locate the bucket where the incomplete multipart upload to delete resides, and click the bucket name to go to the bucket management page.
4. Click File List on the left sidebar.
5. Click Incomplete Multipart Upload.
6. View incomplete multipart uploads on the page.
7. Click Delete on the right of an incomplete multipart upload to delete it. You can also click Clear Incomplete Multipart Uploads at the top to quickly delete all incomplete multipart uploads.

After you perform the Clear Incomplete Multipart Uploads or Delete operation, the list will be empty.

Clearing incomplete multipart uploads regularly with lifecycle policy

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Locate the bucket where the incomplete multipart upload to delete resides, and click the bucket name to go to the bucket management page.
4. On the left sidebar, click Basic Configurations > Lifecycle to enter the lifecycle management page.


5. Click Add a Rule, and you can see the configuration items as shown below. For example, you can set to delete incomplete multipart uploads across the bucket seven days after creation.


6. Click OK, and you can find this new lifecycle rule in the console.



Searching for Objects

Last updated：2025-11-28 18:09:59

Overview

You can use the COS console or object key to search for uploaded objects and folders.
ObjectKey is the unique identifier of an object in the bucket, which can be understood as the file path. For example, if the object key is doc/picture.jpg, it means the image file picture.jpg is stored under the doc path (or folder) in Cloud Object Storage (COS).
List object keys to search for specific targets by specified object keys. In addition, you can also search for objects by object key prefix, which means specifying part of the object key (such as doc) to find all objects with the same prefix doc.

Applicable Scenarios

Tencent Cloud COS supports listing object keys by prefix. You can also use the / character in object keys to implement a hierarchical structure similar to the traditional file system. COS also supports using a delimiter to select and browse keys hierarchically.
You can list all keys in a single bucket in UTF-8 binary order of prefixes or filter the key list by specifying the prefix. For example, adding the parameter t will list the tencent object, but skip objects prefixed with a or other characters.
Add the / delimiter and object keys will be reorganized based on this delimiter. You can combine prefixes and delimiters to achieve folder-like searching feature.
Tencent Cloud COS supports an unlimited number of objects in a single bucket, so the object list may be very large. For convenient management, a single List Objects request returns up to 1000 objects and includes a return indicator to inform if truncation exists. If truncation exists, it means that there are more objects on the next page. You can trigger multiple List Objects requests based on indicators and the separator to list all object keys or query the necessary content.

Usage

Using the COS Console

The COS console supports prefix search and fuzzy search.
Prefix search: Search based on the same object prefix and show objects and folders with the same object prefix in the search results.
Fuzzy search: Search with custom keywords and show objects and folders containing the keywords in the search results.
Note:
Search object prefixes are case sensitive when searching for objects.
Besides, you can also use the COSBrowser tool to do fuzzy search. For details, see COSBrowser Overview.

Searching Objects by Prefix

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to go to the bucket list page.
3. Find the bucket where the object is located and click the bucket name to enter the bucket management page.
4. In the left sidebar, select File List to go to the file list page.
5. In the search box at the top of the page, input the object name prefix and click

to show objects or folders with the same name prefix in the current bucket. For example, input img/ to display search results for objects or folders with img/.


Using Fuzzy Keywords to Search for Objects

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to go to the bucket list page.
3. Find the bucket where the object is located and click the bucket name to enter the bucket management page.
4. In the left sidebar, select File List to go to the file list page.
5. At the top of the page, enter search keywords in the search box and click

to show search results containing the keyword, including objects and folders. For example, enter keyword t to display objects and folders containing t.


Using the REST API/SDK

You can use REST APIs to trigger a List Objects request. For details, see the GET Bucket (List Objects) API document.
You can directly call the SDK to list objects for searching objects. For details, see the following language SDK documents:
Android, C, C++, .NET, Flutter, Go, iOS, Java, JavaScript, Node.js, PHP, Python, React Native, Mini Program, Harmony.

Using Tools

You can use tools to search for objects, such as COSBrowser, COSCMD, COSCLI. For more tool introductions, see Tool Overview.


Copying Objects

Simple Copy

Last updated：2025-09-25 15:54:53

Use Cases

You can create a copy of an object in Cloud Object Storage (COS) with a simple copy operation. The copy operation creates a copy of a COS object of up to 5 GB in a single request. To copy an object over 5 GB, use the Multipart Copy API. With the copy operation, you can:
Create a copy of an object.
Rename an object by copying it and deleting the original one.
Modify the storage class of an object. You can select the same object key as both the source and target and modify the storage class.
Copy objects across different COS regions.
Modify object metadata. You can select the same object key as both the source and target and modify object metadata.
In the copy operation, the metadata of the source object is inherited by default, while the creation date is subject to the creation date of the target object.
Note:
Unsupported operations for objects in Archive Type (such as ARCHIVE, DEEP ARCHIVE) include initiating a PUT Object - Copy request.
Buckets with MAZ configuration enabled do not support copying MAZ storage type to SAZ storage type.
To copy objects, a sub-account must have these permissions: PutObject, GetObject, and GetObjectACL. For details, see the COS Authorization API.

Usage

Using the COS Console

Note:
Support shortcut keys Ctrl+C and Ctrl+V to copy and paste.
Support shortcut key Shift for multiple selection.
The target path cannot be the same as the source path. Otherwise, the paste will fail.
1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to enter.
3. Find the corresponding bucket and click the bucket name to enter the file list page.
4. Select the object or folder to copy (support multiple selections), click More Actions > Copy File.
5. After the prompt displayed for successful copy, select the target path, click More Actions > Paste to paste.
For example, paste under the target folder in a bucket.


Using REST API/SDK

Use REST API to initiate a copy object request. For details, see PUT Object - Copy API document.
You can call the object copy method setting in the SDK. For details, see the following language SDK documents:
Android, C, C++, .NET, Go, iOS, Java, JavaScript, Node.js, PHP, Python, Mini Program.

Tools

You can directly copy objects using the following tools:
COSBrowser
COSCMD
COSCLI

Multipart Copy

Last updated：2025-11-11 16:58:16

Overview

To copy an object that is greater than 5 GB, you must use multipart copy. First, use the multipart upload API to create an object. Then use the Upload Part - Copy API and specify the x-cos-copy-source header to determine the source object that you want to copy. The process is outlined below:
1. Initialize an object for multipart upload.
2. Copy the data of the source object; specify the x-cos-copy-source-range header to determine how much data to copy at a time (you can specify to copy up to 5 GB at a time).
3. Complete the multipart upload.
Note:
The SDKs provided by COS can be used to easily implement a multipart copy operation.

Directions

Using REST APIs

Use REST APIs to directly initiate a multipart copy request. For more information, see the following API documentation:
Initiate Multipart Upload
Upload Part - Copy
Complete Multipart Upload
Abort Multipart Upload

Using SDKs

Use multipart copy directly through SDKs. For more information, see the SDK documentation for the corresponding programming language below:
Android SDK
C SDK
C++ SDK
.NET(C#) SDK
Go SDK
iOS SDK
Java SDK
JavaScript SDK
Node.js SDK
PHP SDK
Python SDK
Mini Program SDK

Previewing or Editing Object

Last updated：2025-07-16 17:46:43

Overview

You can preview object content and edit text objects online in the COS console. For the file types supported for preview, see File Preview Overview.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Locate the bucket where the object resides and click the bucket name.
4. Click File List on the left.
5. Find the target object and click Preview in the Operation column on the right.
6. In the pop-up window, preview the object.
Note:
For buckets created after January 1, 2024, previewing objects of any type using the COS default domain is not supported. Use Custom Origin Server Domains to preview objects.
Additionally, for document files, you can click to copy the temporary link (valid for one hour) and preview it in your browser.
If it is an audio/video file, it will directly open the player in the browser for playback.
7. For a text file, you can click Edit text in the preview window.
8. Authorize to log in to COSBrowser as prompted to edit the file, then you can proceed with file editing.

Viewing Object Information

Last updated：2024-01-06 15:06:14

Overview

You can view the attributes (such as size and address) and configurations (object access permission, storage class, etc.) of an object in the COS console.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Locate the bucket where the object resides and click the bucket name.
4. Click File List on the left sidebar.
5. Find the target object and click Details in the Operation column on the right. On the object details page, you can view the object size and address, obtain the signed URL, and configure the object.


Sorting and Filtering Objects

Last updated：2024-01-06 15:06:15

Overview

This document describes how to sort/filter objects in the file list of the bucket through the COS console.
Note:
Object sorting/filtering is only available if the number of files/folders is less than 1,000.

Directions

1. Log in to the COS console.
2. Click Bucket List in the left sidebar.
3. Click the name of the desired bucket to view its file list.
4. Click

,

, or

,

to sort/filter objects as needed.
Note:
Currently, sorting by file name/file size/modification time, and filtering by storage class are supported.

Restoring Archived Objects

Last updated：2025-08-18 11:14:20

Overview

You can restore an object from the ARCHIVE or DEEP ARCHIVE storage class through the COS console. The restoration operation will create a copy of the object in the STANDARD storage class. You can read, download, or perform other operations on the copy within its validity period. The copy will be deleted automatically after expiration. For more information about storage classes, see Storage Class Overview.
Note:
If you restore objects from ARCHIVE, the object copy will be billed at STANDARD rates. If you restore objects from DEEP ARCHIVE, the object copy will be charged request fees at the marked unit price, and traffic fees at STANDARD rates. For more information, see Product Pricing.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Locate the bucket where the object resides and click the bucket name.
4. Click File List on the left sidebar.
5. Locate the object to restore and click Recover in the Operation column. To restore multiple archived objects at a time, select them all in the object list, and click Recover from the More Actions drop-down list at the top.
6. In the pop-up window, configure the restoration mode and the validity period (in days) of the copy.

The parameters are described as follows:
Restoration Mode: Expedited, Standard, or Bulk retrieval.
Expedited retrieval: This is the fastest mode. Archived files can be restored within 1-5 minutes. If you need to access your archival data urgently, you can use this mode to greatly reduce the restoration time. Please note that the expedited retrieval mode is not available for DEEP ARCHIVE.
Standard retrieval: Files can be restored from ARCHIVE within 3-5 hours, and from DEEP ARCHIVE within 12 hours.
Bulk retrieval: This is the lowest-cost mode. If your need for the archival data is not urgent, this mode can usually restore massive amounts of data from ARCHIVE within 5-12 hours, and from DEEP ARCHIVE within 48 hours, both with an ultra-low cost.
Note:
The QPS of data restoration requests is limited to 100.
The data retrieval fees vary by mode; for details, please refer to Data Retrieval Fees.
Validity: The number of days after which the copy will automatically expire and be deleted. The value range is 1 to 365 days. After the object is successfully restored, you can click OK again to change the validity period of the copy in the pop-up window.
7. Click OK. The object enters the restoration process. During this process, you can check the object recovery progress in the following two ways:
7.1 Refresh the page file list to view the recovery status and expected recovery time of the object in the storage type column.

7.2 You can click Details to go to the object details page to check the restoration progress.

8. Once the object has been successfully restored, you can access, download, and modify the storage type of the object from its details page.
Note:
After the object recovery is complete, before deleting the expired copies, Modifying Storage Class (e.g., to standard storage). The original object type will be replaced from archive storage to standard storage type.

9. If you need to change the validity period of the copy, follow step 5 and click Recover again to modify the settings in the pop-up window.



Exporting Object URLs

Last updated：2024-01-06 15:06:14

Overview

The COS console allows you to export the URLs of objects in a bucket, including objects that have been uploaded or in a folder. Batch export is supported.
Note:
You can export up to 100 object URLs at a time, meaning you can select a maximum of 100 objects. If you need to export more than 100 object URLs, you can use APIs or SDKs. For more information, please see GET Bucket (List Objects).

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Locate the bucket where the object resides and click the bucket name.
4. Click File List on the left sidebar.
5. Select the desired objects, click More Actions > Export URLs.

6. In the pop-up window, configure the following information:

Protocol: HTTPS is selected by default. You can select HTTPS or HTTP as needed. Specified Domain: the default origin domain is supported. Expiration Time: The default value is 60 minutes. You can set it to an integer ranging from 5 to 60.
7. Click Confirm to export the object URLs.
8. In the exported file, you can view the object names and object URLs. Note that you need to view the URLs within the validity period.

Creating a File

Last updated：2025-11-07 15:12:00

Introduction

You can create an empty file through the COS console.
Note:
Character up (↑): CAN (24)
Character down (↓): EM (25)
Character right (→): SUB (26)
Character left (←): ESC (27)

Prerequisites

Bucket created.

Operation Steps

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to go to the bucket list page.
3. Find the bucket where you need to create a file, click its bucket name, and enter the bucket management page.
4. In the left sidebar, select File List to go to the file list page.
5. Hover over "Create Folder" to automatically show a dropdown with "Create File". Click Create a file.

6. In the pop-up window, enter Folder Name and click OK.

Note:
Upon completion, COS creates an empty file with a size of 0 and a storage type of standard storage by default.


Folder Management

Creating Folder

Last updated：2024-01-06 15:06:14

Overview

COS stores objects in a flat structure with no traditional folder concept. In order to make COS customary, we turn an object into a "folder" by suffixing it with / in its key. In fact, a "folder" in COS is an object with a storage capacity of 0 KB. For more information, see Folder and directory.
Note:
The length of the folder name cannot exceed 255 characters, and the following ASCII control characters are not supported:
Up (↑): CAN (24)
Down (↓): EM (25)
Right (→): SUB (26)
Left (←): ESC (27)

Prerequisites

A bucket is created. For operation details, see Creating a Bucket.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Locate the bucket for which a folder is to be created, and click the bucket name to go to the bucket management page.
4. Click File List on the left sidebar.
5. Click Create Folder.
6. In the pop-up window, enter the folder name and click OK.

Note:
Folders cannot be renamed. Please exercise caution when naming folders.
Folder naming rules are as follows:
The folder name can contain digits, letters, and visible characters.
A subdirectory can be created quickly by separating the path by a slash (/).
Do not start with/ or use two or more / consecutively.
Do not leave the folder name empty.
Do not use .. as the folder name.

Deleting Folder

Last updated：2024-01-06 15:06:14

Overview

You can delete a created folder in the COS console.
Note:
Deleting a folder in the console will delete the folder and all objects in it. Please proceed with caution.

Directions

1. Log in to the COS console.
2. In the left sidebar, click Bucket List to go to the bucket list page.
3. Locate the bucket where the folder to delete resides, and click the bucket name to go to the bucket management page.
4. In the left sidebar, choose File List to go to the file list page.
5. Locate the folder to delete, and click Delete in the Operation column.
6. In the pop-up window, click OK.

Sharing Folder

Last updated：2024-01-06 15:06:14

Overview

You can share a created folder using the COS console.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of a bucket where the folder to share is stored.
4. Click File List on the left sidebar.
5. Find the folder to share, and click More > Share Folder in the Operation column.
6. In the pop-up window, configure Permission, Validity Period, and Password. Then, click OK.
7. You can now view the generated share link in the pop-up window. Click Copy Link and Password and share it with others.
8. The recipient can enter the password and view the shared folder as well as the files in it.
9. On the share file page, the recipient can download the file, or view information such as the object storage class, size, and modification time.

Viewing Folder Details

Last updated：2024-01-06 15:06:14

Overview

In the COS console, you can view the details about folders, including the number and size of objects in them. This document describes how to view folder details.
Note:
COS stores objects in a flat structure with no traditional folder concept. To make COS easy to use, we use objects whose object keys are suffixed with / as folders, but actually a "folder" in COS is an object occupying 0 KB.
If you want to query the number and size of all objects in the current bucket, you can get an inventory report with the inventory feature. For more information, see Inventory Overview.
If the total number of objects in the current folder exceeds 10,000, we recommend that you use the inventory feature to count the objects. For more information, see Inventory Overview.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Locate the bucket where the folder is located and click the bucket name to go to the bucket management page.
4. Click File List on the left.
5. Locate the folder whose details you want to view, and click Statistics in the Operation column.
6. In the pop-up window, you can view the statistical information of the folder, including the total number and size of objects.
Note:
Nested folders inside a folder are counted as objects.
Example 1: If a folder contains 1 empty folder and 5 objects, the total number of objects in the folder will be 6.
Example 2: If a folder contains 1 folder (containing 2 objects) and 5 objects, the total number of objects in the folder will be 8.

Setting Folder Permissions

Last updated：2024-06-03 11:14:27

Overview

You can set access permissions for folders through the Cloud Object Storage (COS) console, allowing users to perform specified operations on them. We recommend that you follow the principle of least privilege when configuring permissions to protect your data assets.
Note:
COS stores objects in a flat structure with no traditional folder concept. In order to make COS customary, we turn an object into a "folder" by suffixing it with / in its key. In fact, a "folder" in COS is an object with a storage capacity of 0 KB.
Folder permissions are essentially access permissions at the object level, which take precedence over the bucket access permissions. Unlike object access permissions, folders support setting public read and write. For more information about object access permissions, refer to Access Permission Types in the object overview.

Directions

1. Log in to the COS console.
2. In the left sidebar, click Bucket List to go to the bucket list page.
3. Locate the bucket where the folder is located and click the bucket name to go to the bucket management page.
4. In the left sidebar, choose File List to go to the file list page.
5. Locate the folder for permission setting, and click Permission.
6. In the pop-up window, set folder permissions as required.Permission settings are described as follows:
Permission Type
Parameter
Description
Public permission
Inherit
Default value, consistent with the bucket permission.
Private Read/Write
Only the root account has the read and write permissions on the folder. Non-root accounts (sub-accounts, the root accounts of other users, or anonymous users) cannot access the folder.
Public Read/Private Write
The root account has the read and write permissions on the folder. Non-root accounts (sub-accounts, the root accounts of other users, or anonymous users) can read content in the folder but cannot write new data into the folder.
Public Read/Write
Both the root account and non-root accounts (sub-accounts, the root accounts of other users, or anonymous users) have the read and write permissions on the folder.
User ACL
User Type
`Root account` indicates the root account of other users. `Sub-account` indicates the sub-account under the current root account.To grant the access permission to a sub-account of another root account, you need to grant the access permission to that root account first and then grant the access permission to the sub-account from that root account.
Read
Permission to read data
Write
Permission to write data
Read ACL
Permission to read folder permission configuration. With this permission, you can obtain folder permission configuration details.
Write ACL
Permission to modify folder permission configuration. With this permission, you can modify folder permission configuration details. Exercise caution with this configuration because it will cause permission changes.
Full control
Includes the Reads, Write, Read ACL, and Write ACL permissions. Exercise caution with this configuration because it grants a wide range of permissions.
7. Click Save in the Operation column.
8. Click Close.

Custom Headers

Last updated：2025-11-18 16:52:54

Feature Overview

An HTTP header (metadata header) of an object is a string sent by the server over HTTP before it sends HTML data to the browser. By modifying HTTP headers (metadata headers), you can modify how the webpage responds as well as certain configurations, such as caching time. Modifying an object's HTTP headers does not modify the object itself.
Note:
For objects of ARCHIVE and DEEP ARCHIVE, custom headers can only be set during upload. Setting custom headers for objects that have already been uploaded is not supported.

How It Works

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Locate the bucket where the object resides and click the bucket name.
4. Click File List on the left sidebar.
5. Locate the object for which a header is to be customized and click More > Custom Header in the operation column on the right. To customize headers for multiple objects, select multiple objects and click More Actions > Custom Header on the top.
6. In the pop-up window, select the parameter type of the metadata header to be set, enter the metadata value, and click OK. COS provides the following six types of object HTTP headers for configuration:
HTTP Header
Description
Example
Content-Type
MIME information of the file
image/jpeg
Cache-Control
File caching mechanism
no-cache: Indicates that cache cannot be used directly, but whether the object has been updated must be verified first with the server. If the object has been updated, it indicates that the cache is expired, and the object must be re-downloaded from the server; if the object has not been updated, the cache is not expired, and the local cache will be used.
max-age=200: Indicates the relative expiration time of the cached content, in seconds.
Content-Disposition
Extension of MIME protocol
inline: Directly previews the content of the file.
attachment: Downloads to the specified path in the browser with the original filename.
attachment; filename="FileName": You should download it to the specified path in the browser with a custom filename. FileName defines the name of the file after download, for example, example.jpg. If the attachment name is in Chinese, it needs to be URL encoded, for example, attachment; filename* = UTF-8' '%E4%B8%AD%E6%96%87.txt
Content-Encoding
File encoding format
gzip
Note: If the Content-Encoding in the header has been modified as gzip, but the file has not been compressed with gzip, a decoding error will occur.
Expires
Expiration date of the cache
Wed, 21 Oct 2015 07:28:00 GMT
x-cos-meta-[custom suffix]
User-defined content
x-cos-meta-via: homepage
Note: Header names support only hyphens (-), digits, and English letters (a-z). Uppercase letters in English will be converted to lowercase letters; other characters, including underscores (_), are not supported.

License request example

Assume that a bucket named "examplebucket-1250000000" was created under account APPID 1250000000, and an object "exampleobject.txt" was uploaded to the bucket's root directory.

Non-custom HTTP header of objects

The sample below shows the headers returned for a request to download this object through a browser or client if no custom HTTP headers are specified.

Request

GET /exampleobject.txt HTTP/1.1
Host: examplebucket-1250000000.cos.ap-beijing.myqcloud.com
Date: Fri, 10 Apr 2020 09:35:16 GMT
Authorization: q-sign-algorithm=sha1&q-ak=AKID8A0fBVtYFrNm02oY1g1JQQF0c3JO****&q-sign-time=1586511316;1586518516&q-key-time=1586511316;1586518516&q-header-list=date;host&q-url-param-list=&q-signature=1bd1898e241fb978df336dc68aaef4c0acae****
Connection: close

Response

HTTP/1.1 200 OK
Content-Type: text/plain
Access-Control-Allow-Origin: *
Last-Modified: Fri, 10 Apr 2020 09:35:05 GMT

Custom HTTP header of objects

Refer to directions to add the following custom header configuration:
Parameter
Value
Content-Type
image/jpeg
Cache-Control
no-cache
Content-Disposition
attachment; filename* = UTF-8' '%E4%B8%AD%E6%96%87.jpeg
x-cos-meta-md5
1234
Then the headers returned for new requests will be as follows:

Request

GET /exampleobject.txt HTTP/1.1
Host: examplebucket-1250000000.cos.ap-beijing.myqcloud.com
Date: Fri, 10 Apr 2020 09:35:16 GMT
Authorization: q-sign-algorithm=sha1&q-ak=AKID8A0fBVtYFrNm02oY1g1JQQF0c3JO****&q-sign-time=1586511316;1586518516&q-key-time=1586511316;1586518516&q-header-list=date;host&q-url-param-list=&q-signature=1bd1898e241fb978df336dc68aaef4c0acae****
Connection: close

Response

HTTP/1.1 200 OK
Cache-Control: no-cache
Content-Type: image/jpeg
Content-Disposition: attachment; filename* = UTF-8' '%E4%B8%AD%E6%96%87.jpeg
x-cos-meta-md5: 1234
Access-Control-Allow-Origin: *
Last-Modified: Fri, 10 Apr 2020 09:35:05 GMT
It can be seen that when custom headers are used, the response format of the page can be changed.

Data Management

Lifecycle Management

Lifecycle Overview

Last updated：2024-03-25 15:28:24

Overview

COS supports object-based lifecycle configuration. You can use lifecycle rules to define operations to perform on applicable objects.
Note:
Up to 1,000 lifecycle rules can be added for each bucket.

Use Cases

Log record

With lifecycle configured, logs stored on COS can be automatically archived after 30 days or deleted after 2 years.

Hot/Cold data tiering

Hot data is frequently accessed in a short period of time after the upload, but then is less or not accessed at all after some time. Therefore, you can set the lifecycle rules to store the data uploaded 30 days ago in the STANDARD_IA storage class and data uploaded 60 days ago in ARCHIVE. This process is called data transition.

Archive management

When you use COS for file archive management, you need to save all historical versions of files for a long time according to the compliance requirements in financial, medical, and other industries. In this case, you can configure the lifecycle to transition and store the historical versions of files in the ARCHIVE storage class.

Configuration Items

You need to configure the following elements to create a lifecycle rule:

Objects

The data to be hit during lifecycle execution. You can customize the lifecycle's scope and data types covered within the scope. During lifecycle execution, the specified scope will be scanned, and the operation will be performed on the configured data types within the scope. You can specify the scope according to the following rules:
Specify by prefix: Data can be matched by directory name or filename prefix.
Specify by tag: Data can be filtered by tag.
The following data types can be configured:
Files on the current version: Files on the latest version in the bucket.
Files on historical versions: Objects on historical versions stored after versioning is enabled. For more information on versioning, see Overview.
Delete marker: Marker indicating that the object has been deleted. The lifecycle feature can remove the marker automatically after all historical versions are deleted. For more information on the delete marker, see Delete Markers.
Incomplete multipart uploads: Fragments generated due to incomplete multipart uploads.

Operations

The operation to be performed when the object is hit:
Data transition: transitions objects to STANDARD_IA, INTELLIGENT TIERING, ARCHIVE, or DEEP ARCHIVE after a specified period.
Expiration: Deletes objects after their specified expiration time.

Time

The time condition for triggering the above operations: Based on number of days: Specify after how many days since an object was last modified that an action will occur.

Notes

Note:
For how to use the lifecycle, see Configuring Lifecycle.
File modification time Lifecycle supports triggering rule execution based on the object modification time. Only file write operations such as PUT Object, PUT Object - Copy, POST Object, and Complete Multipart Upload APIs will update the object modification time. The modification time of objects transitioned based on the lifecycle won't be updated.
For example, you configured a lifecycle rule at 15:00 on the 1st day of the month to delete files one day or longer after they are modified. Then, at 00:00 on the 2nd day, the lifecycle task scans for files that were modified over one day ago and deletes them. Files uploaded on the 1st day will not be deleted at 00:00 on the 2nd day, as the time elapsed since their modification is less than one day. Instead, these files will be deleted at 00:00 on the 3rd day.
Note:
A lifecycle can be set to as long as 3,650 days.

Effective time

Data transition

Supported regions

Data transition is supported in public cloud regions. Finance Cloud regions support only data transition to the STANDARD_IA storage class.

One-way transition

Data transition is one-way (STANDARD > STANDARD_IA > ARCHIVE, or STANDARD > ARCHIVE) and cannot be done reversely. You can only call PUT Object - Copy (for non-ARCHIVE/DEEP ARCHIVE only) or POST Object restore (for ARCHIVE and DEEP ARCHIVE only) to restore data from a colder storage class to a hotter one.

Eventual consistency

If multiple rules are configured for the same set of objects and conflict with each other (excluding the configuration for deletion upon expiration), COS will execute the rule that transitions the objects to the coldest storage class.
For example, if rules A and B are configured to transition the objects to STANDARD_IA and ARCHIVE 90 days after file modification respectively and both hit the same object test.txt, rule B will be executed.
Rule
Resource
Operation
Time Condition
Execution
Rule A
test.txt
Transition the object to the STANDARD_IA class.
90 days after file modification
Execution fails due to a rule conflict.
Rule B
test.txt
Transition the object to the ARCHIVE class.
90 days after file modification
Execution succeeds.
Note:
It is strongly recommended that you do not configure conflicting lifecycle rules for the same set of objects in COS, because this may result in additional fees.
Transitioning an object won't change the time the object was uploaded or modified.

Deletion upon expiration

How it works

When an object matches the specified lifecycle rule of deletion upon expiration, Tencent Cloud will add the object to the asynchronous deletion queue. There may be a certain delay in actual deletion. You can get the current status of the object by performing GET or HEAD Object operation.

Eventual consistency

If multiple conflicting rules are configured for the same set of objects, the rule that will expire first will be performed first. Deletion upon expiration will be performed before the transition.
For example, if rules C and D are configured to transition the objects to STANDARD_IA and delete the objects 180 days after file modification respectively and both hit the same object test.txt, rule D will be executed.
Rule
Resource
Operation
Time Condition
Execution
Rule C
test.txt
Transition the object to the STANDARD_IA class.
180 days after file modification
Execution fails due to a rule conflict.
Rule D
test.txt
Delete the object.
180 days after file modification
Execution succeeds.
Note:
It is strongly recommended that you do not configure conflicting lifecycle rules for the same set of objects in COS, because this may result in different fees.

Cost considerations

Execution instructions

When the lifecycle feature performs a deletion operation, a backend deletion request will be generated. When it performs a transition operation, backend deletion and write requests will be generated. The requests generated by the above operations will be included in the request bill. For example, transitioning the file test.txt in STANDARD storage class to STANDARD_IA through the lifecycle will generate two requests to delete the data in STANDARD and write the data in STANDARD_IA.
When an accident occurs or there are too many objects in the bucket, lifecycle execution may fail. For failures due to other reasons, perform the GET or HEAD Object operation to get the current object status. Tencent Cloud cannot provide an accurate bill unless the lifecycle execution is completed.

Time-insensitive

Please note that the minimum storage duration is 30 days, 90 days, and 180 days for the STANDARD_IA/INTELLIGENT TIERING, ARCHIVE, and DEEP ARCHIVE storage classes, respectively. No additional storage fees will be incurred for the transition or delete action itself. COS will ignore lifecycle configurations less than 30/90/180 days, and perform only correct configurations upon your request.
For example, if an object in STANDARD_IA is transitioned before 30 days, it will start incurring ARCHIVE storage fees on the transition day and continue incurring STANDARD_IA storage fees until the 30th day. Another example is that if an archived object is deleted upon expiration before it is stored for 90 days, it will continue incurring ARCHIVE storage fees until the 90th day. It works the same way with DEEP ARCHIVE.

Size limits

There is a minimum size limit for objects in the STANDARD_IA, INTELLIGENT TIERING, ARCHIVE, and DEEP ARCHIVE storage classes. For example, if an object smaller than 64 KB is uploaded to the STANDARD_IA storage class, it will be calculated as 64 KB. To reduce user costs, lifecycle execution does not transition the storage classes of objects that are smaller than 64 KB.
Note:
Lifecycle execution does not transition the storage classes of objects that are smaller than 64 KB.

Lifecycle Configuration Elements

Last updated：2024-03-25 15:28:24

Basic Structure

A lifecycle configuration is specified as XML and consists of one or more lifecycle rules. The structure is as follows:
<LifecycleConfiguration>
    <Rule>
        <ID>**your lifecycle name**</ID>
        <Status>Enabled</Status>
        <Filter>
            <And>
                <Prefix>projectA/</Prefix>
                <Tag>
                    <Key>key1</Key>
                    <Value>value1</Value>
                </Tag>
            </And>
        </Filter>
        **transition/expiration actions**
    </Rule>
    <Rule>
        ...
    </Rule>
</LifecycleConfiguration>
Each rule contains the following:
ID (optional): Indicates the content of the rule and is customizable.
Status: Indicates whether the rule is enabled or disabled.
Filter: Identifies objects to which the rule applies.
Action: Specifies actions (operations) that need to be performed on objects that match the above description.
Time: Days (after the object's last modified date) or Date based on which actions are performed on objects.

Rule Description

Filter element

For all objects in bucket

Specify an empty filter, in which case the rule applies to all objects in the bucket.
<LifecycleConfiguration>
    <Rule>
        <Filter>
        </Filter>
        <Status>Enabled</Status>
        **transition/expiration actions**
    </Rule>
</LifecycleConfiguration>

Based on object key prefix

Specify a prefix-based filter so that the rule applies only to objects with the specified prefix. For example, you can filter out objects by the prefix logs/:
<LifecycleConfiguration>
    <Rule>
        <Filter>
           <Prefix>logs/</Prefix>
        </Filter>
        <Status>Enabled</Status>
        **transition/expiration actions**
    </Rule>
</LifecycleConfiguration>

Based on object tag

Specify a filter based on key and value so that the rule applies only to objects with the specified tag. For example, you can filter out objects by tags key=type and value=image:
<LifecycleConfiguration>
    <Rule>
        <Filter>
           <Tag>
              <Key>type</Key>
              <Value>image</Value>
           </Tag>
        </Filter>
        <Status>Enabled</Status>
        **transition/expiration actions**
    </Rule>
</LifecycleConfiguration>

Using multiple filters

In COS, you can combine multiple filters by using AND. For example, you can filter out objects by the prefix logs/ as well as tags key=type and value=image:
<LifecycleConfiguration>
    <Rule>
        <Filter>
            <And>
                <Prefix>logs/</Prefix>
                <Tag>
                      <Key>type</Key>
                      <Value>image</Value>
                </Tag>
            </And>
        </Filter>
        <Status>Enabled</Status>
        **transition/expiration actions**
    </Rule>
</LifecycleConfiguration>

Action element

Specify one or more predefined actions in a lifecycle rule so that these actions will be performed when any objects meet the rule.

Transition action

Specify the Transition action to transition objects from one storage class to another. For a versioning-enabled bucket, this action applies to the current object version. You can set the transition date to as short as 0 days. For example, you can transition objects to the ARCHIVE storage class in 30 days:
<Transition>
    <StorageClass>ARCHIVE</StorageClass>
    <Days>30</Days>
</Transition>

Deletion upon expiration

Specify the Expiration action to delete expired objects. For a versioning-disabled bucket, expired objects will be permanently deleted. For a versioning-enabled bucket, expired objects are added with a DeleteMarker and become the current version. For example, you can delete objects that expire in 30 days:
<Expiration>
    <Days>30</Days>
</Expiration>

Incomplete multipart upload

Note:
You cannot specify object tags and clear incomplete multipart uploads at the same time in the same lifecycle rule.
Specify the AbortIncompleteMultipartUpload action to delete multipart upload tasks with the specified UploadId if they are not successfully completed within the predefined time period. Then, these tasks cannot be resumed or indexed. For example, you can abort multipart upload tasks not successfully completed within 7 days:
<AbortIncompleteMultipartUpload>
   <DaysAfterInitiation>7</DaysAfterInitiation>
</AbortIncompleteMultipartUpload>

Non-current object

In a versioning-enabled bucket, the Transition action only applies to objects of the latest version, and the Expiration action only results in adding delete markers to the objects. Therefore, COS provides the following actions for objects of non-current versions:
Specify the NoncurrentVersionTransition action to transition non-current objects to another storage class at the specified time. For example, you can transition non-current objects to the ARCHIVE storage class in 30 days:
<NoncurrentVersionTransition>
    <StorageClass>ARCHIVE</StorageClass>
    <Days>30</Days>
</NoncurrentVersionTransition>
Specify the NoncurrentVersionExpiration action to delete non-current objects at the specified time. For example, you can delete non-current objects in 30 days:
<NoncurrentVersionExpiration>
    <Days>30</Days>
</NoncurrentVersionExpiration>
Specify the ExpiredObjectDeleteMarker action to clear the remaining delete markers. You can set it to true or false. This action is triggered by the NoncurrentVersionExpiration action. If this feature is enabled, when lifecycle execution deletes the last historical version in the lifecycle of an object, COS will automatically delete the remaining delete markers.
The details are as follows. If the NoncurrentVersionExpiration action is in effect:
If there is only one delete marker left, COS will automatically delete the last delete marker regardless of whether ExpiredObjectDeleteMarker is enabled.
If there are multiple delete markers left, COS will check whether ExpiredObjectDeleteMarker is enabled (true), and if so, COS will automatically delete the remaining delete markers; otherwise, COS will not delete them.
<ExpiredObjectDeleteMarker>
    <Days>true|false</Days>
</ExpiredObjectDeleteMarker>

Time element

Based on number of days

Days refers to the number of days since an object was last modified.
For example, if an object is set to be transitioned to the ARCHIVE storage class in 0 days, when the object is uploaded at 2018-01-01 23:55:00 GMT+8, it will be added to the transition queue at 2018-01-02 00:00:00 GMT+8 and transitioned before 2018-01-02 23:59:59 GMT+8.
For example, if an object is set to be deleted upon expiration in 1 day, when the object is uploaded at 2018-01-01 23:55:00 GMT+8, it will be added to the expiration queue at 2018-01-03 00:00:00 GMT+8 and deleted before 2018-01-03 23:59:59 GMT+8.

Based on specific date

Perform the predefined action on all eligible objects based on the filter on the specified Date. The date must be in the ISO 8601 format, and the time is always 00:00 GMT+8. For example, use 2018-01-01T00:00:00+08:00 for January 1, 2018.

Configuring Lifecycle

Last updated：2025-12-03 11:50:24

Scenarios

With lifecycle configuration, the predefined action can be automatically performed when a rule applies to objects. For example:
Transition storage class: After a specified period, objects can be automatically transitioned to different storage classes, such as Infrequent Access Storage (STANDARD_IA), Intelligent Tiering Storage (INTELLIGENT_TIERING), Archive Storage (ARCHIVE), and Deep Archive Storage (DEEP_ARCHIVE).
Expiration: Deletes objects after their specified expiration time.
For more information, see Lifecycle Overview and Lifecycle Configuration Elements.

How to Use

Using the COS Console

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to enter.
3. Find the bucket to enable the lifecycle function, click its bucket name, and enter the Bucket Details Page.
4. 
Select
the basic configuration > lifecycle configuration item on the left. Configuration items are described below.

Access Tracking: If you wish to set a lifecycle rule based on "last access time", enable this option. When enabled, the activation time for access tracking will be the last access time for all objects in a bucket. You may configure the policy type as Last Access Time in Step 5.
Note:
Access Tracking is an allowlist feature and requires allowlisting before usage. Please contact your business representative or contact us. It will take effect on the same day after allowlisting.
Access Tracking feature usage restriction statement is as follows:
Currently only support enabling access tracking for buckets in Beijing, Shanghai, Guangzhou, Beijing Zone 1, Chongqing, Singapore, and Nanjing.
OFS integrated buckets do not support enabling access tracking.
5. Click Add Rule. Configuration items are described.

Basic information
Rule Name : Enter a name for your lifecycle rule.
Policy Type: Divided into last modification time and last access time. This configuration item is only applicable to users added to the access tracking allowlist.
Last Modification Time: Set rules based on the last modification time of objects.
Last Access Time: Set rules based on the last access time of objects. This configuration item requires allowlisting and enabling the access tracking feature (can be enabled in Step 4). Currently only supports downgrading to infrequent access storage.
Note:
After enabling access tracking, COS will start recording the access time of objects in the bucket. All access behaviors before enabling access tracking will not be included in access time analysis.For example, daily statistics on April 1 found that object a (standard storage) had not been accessed for 10 days. Access tracking was enabled that day, and a lifecycle rule was configured as "Tiering to low-frequency storage 5 days after the last access time of object a". Object a will be archived to low-frequency storage on April 6, not April 2.
Scope: This lifecycle rule can apply to the entire bucket or specific objects in the bucket.
When selecting the specified scope, configure at least one of the following:
Object Prefix: You can specify objects with the same object key prefix to execute the lifecycle rule. Regular expressions are not supported.
Object Tag: You can specify objects with the same tag to execute the lifecycle rule. Multiple tags are supported. Note that it is case-sensitive. If you choose the policy type based on last access time, this item cannot be configured.
Note:
You can specify both object prefix and object tag. The relationship between object prefix and object tag, or between object tags, is "AND" (all conditions must be satisfied simultaneously). For example, if you specify the object prefix as "doc" and the tag key-value pair as group = IT in the lifecycle rule, the specified object range includes all objects in the current bucket with the object key prefix "doc" and the object tag group = IT.
When selecting the entire bucket, you can choose whether to configure the exclusion range, which is off by default. Settings can be done based on the following configuration items:
Object Prefix: You can exclude objects with the same file prefix from executing the lifecycle rule. Currently, only one exclusion prefix can be configured. Note that the current exclusion range configuration does not support fusion buckets.

Rule configuration
Select and configure the following information based on the selected policy type.
Last modification time.

Manage current version files: You can enable the option to manage current version files to transition or delete current version objects. Objects in the bucket can be transitioned from STANDARD storage (hot data) to STANDARD_IA storage (cold data), or deleted after expiration.
The storage types from hot to cold are: STANDARD > STANDARD_IA > INTELLIGENT_TIERING > ARCHIVE > DEEP_ARCHIVE. Storage type conversion can only be performed from hot to cold, not in reverse. For an introduction to storage types and region descriptions, see Storage Class Overview.The time is calculated based on the file's modification time in COS. Modifying an object is equivalent to re-uploading the object.
Note:
For buckets with multi-AZ configuration enabled, the lifecycle transition order only supports Standard Storage (Multi-AZ) > Infrequent Storage (Multi-AZ) > Intelligent Tiering Storage (Multi-AZ)/Archive Storage (Multi-AZ).
Manage Files of Historical Versions: You can enable this option to transition or delete historical version objects. If this option is not enabled, we will only process the latest version objects by default.
Note:
The transition and deletion of earlier versions are calculated based on the time when the object becomes an earlier version, not the upload time of the earlier version.
Clean up deletion markers without historical versions: To clean up deletion markers without historical versions, enable the following two items simultaneously.
Enable the "Manage Files of Historical Versions" option and set expiration and deletion for historical versions.
2.Enable the "Clean up deletion markers without historical versions" option.
Note:
This option takes effect depending on the rule for cleaning up historical versions. When you check this option, if the latest version of an object is a deletion marker and the last historical version object is deleted via lifecycle, the remaining multiple deletion markers will be automatically cleaned up on the second day. For example, after the historical version cleanup is completed on January 1, the deletion markers will be automatically cleaned up on January 2. For details, please refer to ExpiredObjectDeleteMarker.
This option cannot be used with expiration and deletion in Managing the current version files simultaneously.
Delete Incomplete Multipart Uploads: When uploading a file, the upload fails due to various reasons, and only a part of it is transmitted. For such damaged files, you can set periodic deletio
Last access time

Manage Version Files: You can enable this option to determine whether to tier version objects by specifying the file's consecutive access days and access count below a certain threshold. Objects in the bucket can be tiered from the standard type (such as Standard Storage) to the infrequent type (such as Infrequent Storage). Once enabled, both current and earlier versions will be tiered.
6. After confirming the information is correct, click OK, and you will see the lifecycle rule.

Note:
After setting a lifecycle rule, repeated modifications may affect its execution. For details on rule effectiveness, see explanation of rule effective time.
7. If you need to stop a lifecycle rule, click Edit, change the status of the corresponding rule to Close or directly delete the lifecycle rule.
8. If you need to clear all lifecycle rules in the current bucket, click Clear All Rules.

Rule Execution Priority

Each bucket can add up to 1,000 lifecycle rules. If multiple rules are configured for the same group of objects and conflicts exist, COS will execute the lifecycle rules according to the following priorities based on different types of conflicts. For more detailed information, please refer to Lifecycle Overview - Configuration Elements - Operation.
Note:
Tencent Cloud COS strongly reminds you not to configure multiple lifecycle rules with conflicting conditions for the same set of objects, as conflicting executions may lead to different expense performances.

Different Operations of the Same Lifecycle Rule

If you configure a lifecycle rule and set different operations (such as transition and deletion) for the same group of objects in the rule, the execution rules and examples for these operations are as follows:
Execute the Rule
Example
If both deletion and transition operations are applicable at the same time, the deletion operation takes precedence.
Rule R1:
1. The file test.txt will transition to infrequent access storage after being modified for 90 days.
2. The file test.txt will be deleted 90 days after modification.
Expected outcome:
Priority execution for 2, the file test.txt will be deleted after 90 days; execution of 1 failed.
If multiple deletion operations are applicable at the same time, prioritize the shorter time deletion operation.
Rule R1:
1. The file with the specified prefix a will be deleted after 180 days.
2. Files with the specified prefix aa will be deleted after 90 days.
Expected result:
Assume that the bucket has the file aaa.png, specifying the prefix a and specifying the prefix aa both hit the same file aaa.png, prioritize 2, and the file aaa.png is deleted after 90 days; 1 fails to execute.
If multiple transition operations are met at the same time, prioritize the transition operation with colder target storage type.
Rule R1:
1. The file test.txt will transition to infrequent access storage after being modified for 90 days.
2. The file test.txt will transition to archive storage 90 days after modification.
Expected outcome:
Priority execution for 2, the file test.txt will be transition to archive storage after 90 days of modification; execution of 1 failed.

Different Operations For Different Lifecycle Rules

If you configure multiple lifecycle rules and set different operations (such as transition and deletion) for the same group of objects in different rules, the execution rules and examples for these operations are as follows:
Execute the Rule
Example
If multiple different deletion and transition operations among rules are applicable at the same time, prioritize the shortest deletion time operation.
Rule R1:
1. The file test.txt will transition to infrequent access storage 50 days after modification.
2. The file test.txt will be deleted 10 days after modification.
Rule R2:
1. The file test.txt will transition to infrequent access storage 10 days after modification.
2. The file test.txt will be deleted 30 days after modification.
Expected outcome:
Match rule R1, prioritize executing 2; the file test.txt is deleted 10 days after modification; if rule R2 execution fails, the execution of 1 in R1 also fails.
If multiple deletion operations among different rules are applicable at the same time, prioritize the shorter time deletion operation.
Rule R1: Delete the file test.txt 10 days after modification.
Rule R2: Delete the file test.txt 30 days after modification.
Expected outcome:
Match rule R1; the file test.txt is deleted 10 days after modification; rule R2 execution fails.

Rule R3: Delete the file example.txt 10 days after modification.
Rule R4: Delete the file example.txt 10 days after modification.
Rule R5: Delete the file example.txt 50 days after modification.
Expected outcome:
Match rule R3 or R4; the file example.txt is deleted 10 days after modification; rule R5 execution fails.
If multiple rules with different transition operations are satisfied simultaneously, prioritize the transition operation with the colder target storage type.
Rule R1: The file test.txt will be transition to infrequent access storage 10 days after modification.
Rule R2: The file test.txt will be transition to archive storage 10 days after modification.
Expected outcome:
Match rule R2; the file test.txt is transition to archive storage 10 days after modification; rule R1 execution fails.

Rule R3: The file example.txt will be transition to infrequent access storage 10 days after modification.
Rule R4: The file example.txt will be transition to archive storage 10 days after modification.
Rule R5: The file example.txt will be transition to archive storage 10 days after modification.
Expected outcome:
Match rule R4 or R5; the file example.txt is transition to archive storage 10 days after modification; rule R3 execution fails.

Using the REST API

Configure and manage the lifecycle of objects in a bucket through the REST API as described in the following API documentation:
PUT Bucket lifecycle
GET Bucket lifecycle
DELETE Bucket lifecycle

Using SDKs

Directly call the lifecycle method in the SDK. For more information, see the SDK documentation for the corresponding programming language below:
Android SDK
C SDK
C++ SDK
.NET SDK
Go SDK
iOS SDK
Java SDK
JavaScript SDK
Node.js SDK
PHP SDK
Python SDK
Mini Program SDK

Setting Static Website

Last updated：2025-11-18 15:31:02

Concepts

A static website refers to a website that contains static content (such as HTML) or client-side scripts. Users can configure the static website via the COS Console for the bucket bound with custom domains.A dynamic website contains server scripts such as PHP, JSP, or ASP.NET, and needs to be processed on a server. You can host static websites on Tencent Cloud COS, but cannot write server scripts. For deployment of a dynamic website, recommended for use Cloud Virtual Machine (CVM) for server-side code deployment.

Here is an example to introduce the ability to host a static website.
A user created a bucket named examplebucket-1250000000 and uploaded the following files: 
index.html
404.html
403.html
test.html
docs/a.html
images/

Static website

Before enabled: Access the bucket via the following default domain name. When a download prompt pops up, save the index.html file to the local.
https://examplebucket-1250000000.cos-website.ap-guangzhou.myqcloud.com/index.html
After enabled: Access the bucket via the following access node, and then you can view the content of index.html directly in the browser.
https://examplebucket-1250000000.cos-website.ap-guangzhou.myqcloud.com/index.html
Note:
For buckets created after January 1, 2024, direct access to objects through the COS default domain (including static website domain) is no longer supported. For details, see COS Bucket Domain Usage Security Management Notification (Effective January 2024). You can access objects through a custom domain name. For details, see Enable Custom Origin Site Domain.

Forced HTTPS

Before enabled: When the request source is HTTP, the access node URL keeps the HTTP unencrypted transport protocol.
https://examplebucket-1250000000.cos.ap-guangzhou.myqcloud.com/index.html
After enabled: The access node always maintains the HTTPS encrypted transport protocol regardless of whether the request source is HTTP or HTTPS.
https://examplebucket-1250000000.cos-website.ap-guangzhou.myqcloud.com

Index document

An index file, the homepage of the static website, is a page returned when the root directory or any subdirectory of a website is requested, and is usually named index.html. When you access a static website via a bucket access domain name, such as https://examplebucketbucket-1250000000.cos-website.ap-guangzhou.myqcloud.com, and no specific page is requested, the web server will return the homepage.
When your user accesses any directory (including the root directory) in a bucket using a URL ending with /, the index document in that directory will be matched preferentially. / is not mandatory in the root URL, so the index document is returned in response to either of the following URLs.
http://www.examplebucket.com/
http://www.examplebucket.com
Note:
If a folder is created in the bucket, the index file needs to be added at each level of the folder.

Error document

Assume that when you visit the following page before configuring the error document, a 404 status code is returned, and the default error information is displayed on the page.
https://examplebucket-1250000000.cos-website.ap-guangzhou.myqcloud.com/webpage.html
When you visit the following page after configuring the error document, a 404 status code is also returned, but the specific error information is displayed on the page.
https://examplebucket-1250000000.cos-website.ap-guangzhou.myqcloud.com/webpage.html

Redirection rules

Note:
To configure redirection rules for a hosted static website, the path of the replacement file should be an object path in the bucket.
Redirection rule priority is to trigger force HTTPS or prefix match redirection first, and finally trigger error code redirection.

Configure error code redirection

If the webpage.html document is set to Private Read/Write, when a user tries to access it, a 403 error is returned.
After the 403 error code is redirected to 403.html, the browser will return the content of 403.html.
If you do not configure a 403.html document, the browser will return an error document or default error message.



Configure prefix match

Note:
Prefix match does not support wildcards. If you want to redirect two folders prefixed with index1/ and index2/, you cannot use index*/ as the match rule; instead, you should create corresponding match rules separately.
When you rename the docs/ folder to documents/, users will encounter an error when accessing the docs/ folder. So, you can redirect requests with the docs/ prefix to documents/.

When you delete the images/ folder (i.e., all objects with the images/ prefix are removed). You can add a redirection rule to redirect any request for objects with the images/ prefix to the test.html webpage.


Scenarios

You can set up a bucket to host a static website through the COS console, and access the static website via the bucket's static website domain.

Notes

If your static website is not external, it is recommended to set the permission of the bucket to private read/write, and achieve full-link protection of data via CDN origin authentication and CDN authentication. For details, see CDN acceleration configuration.
If you wish to make the static website accessible to everyone, set the bucket permission to public-read/private-write. However, note this may lead to misappropriated resources, so configure with caution. See anti-fraud guidelines to reduce risks of being stolen.
After enabling static website configuration and setting a custom domain, you need to use the custom domain to access the origin server for the changes to take effect. If you use the COS default domain to access, there will be no static website effect (for example, unable to present normal layout).

Prerequisites

The bucket is created. For more information, see the document Create Bucket.

Operation Steps

Enabling Static Website

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to go to the bucket list page.
3. Locate the bucket to be used to host a static website, click its bucket name, and enter the bucket details page.
4. In the left navigation bar, select Permission Management > Bucket Access Permissions, then choose public-read/private-write in the public permission section and save.
5. In the left sidebar, select Basic Configuration > Static Website. In the Static Website section, click Edit and toggle on the Enable button under Status.
6. Fill in the configuration items for the static website in sequence.

The configurations are described as follows:
Access node: When you enable static website, you will GET a static website access node (i.e., static website domain) associated with the bucket. The access node supports only GET/HEAD Object operations.
Note:
The access nodes of a static website are case sensitive. Please note that the file name and suffix English case filled in when configuring the index file, error file, and redirection rule prefix match should be consistent with the files in the bucket.
Force HTTPS (Optional): When you enable force HTTPS, the access node of the static website will forcibly use the HTTPS protocol for user access.
Note:
If you access the static website through a custom domain name while force HTTPS is enabled, confirm the custom domain name has a related certificate configured correctly. If there is a certificate anomaly, it will navigate to the default domain name.
Ignore HTML extension (optional): If the access path is index, the index.html object will be automatically matched and returned.
Index document (required): The index document is the homepage of a static website. It is the returned page when users request the root directory or any subdirectory of a website, usually named index.html.
Note:
If a folder is created in the bucket and web pages reference these folders, add an index document at each folder hierarchy level.
For COS buckets created after January 1, 2024, if you use the static website domain name to access them, the index document will not return webpage content at the time of request but will trigger a direct download.
Error document (optional): An error document refers to the page returned when an error occurs during static website access. This configuration item allows you to define your own error documents. When a static website cannot respond to a user request, it will return a specified custom error page. For example, if you configured an error document named error.html, the webpage will return error.html when a user encounters an HTTP error during access, providing help and guidance. If you have not configured an error document, the page will return the default error message when a user encounters an HTTP error at this time.
Note:
The error document configuration can support files in the root directory or subdirectories of the bucket. Use browser recognizable files like .html or .htm. If you use a file that the browser cannot recognize, such as .zip, most browsers will display errors, be unable to access, or deny the access request.
Error document response code: This item is displayed if an error document is set. You can configure the HTTP response code when the error document is returned to be the original error code or 200.
Redirection rule (optional): With redirection rules, you can conditionally redirect requests based on a specific file path, prefix in request, or response code.
For example, if you delete or rename a certain file in the bucket, you can add one redirection rule to redirect the request to access the file to other files.
Type: Refers to the original type of the redirection rule.
Error code: Currently, only 4xx error codes (such as 404) can be redirected in the redirection rules. You can optionally customize the error page. If a user triggers a corresponding HTTP error, you can provide additional guidance on this error page.
Prefix matching: You can use the prefix matching rules to redirect files or folders in the bucket. For specific examples, see Redirection Rules.
Note:
Prefix match does not support wildcards. If you want to redirect two folders prefixed with `index1/` and `index2/`, you cannot use `index*/` as the match rule. You must create the corresponding match rules separately.
Description: The original error code or prefix of the redirection rule.
Rule: The replacement rule for redirection, including replace path and replace prefix.
If you choose error code as the type, the rule defaults to "replace path". Replace path means the request address will be replaced with the path of the new page when the redirection rule is triggered.
If you choose prefix as the type, the rule supports selecting "replace path" or "replace prefix".
Note:
Replace path must be the object path in bucket.
Replacement content: The file path or prefix to be replaced.
7. After the settings are completed, click Save.

Configuring Custom Domain

Buckets created after January 1, 2024 do not support preview when users access files via the default domain name (including the static website domain). Therefore, after enabling Static Website, you need to configure and use a custom domain name to preview the static website. Directions:
1. Log in to the COS Console, enter the bucket details page, select Domain and Transfer Management, then Custom Origin Server Domains.
2. In the Custom Origin Server Domains section, click Add Domain, input the custom domain name (if not, register a domain name and re-input), and set the origin server type to static website origin server. For details about configuration items, please refer to Enable Custom Origin Site Domain.

3. To make the custom domain name take effect, you need to configure custom DNS parsing to the static website CNAME domain. If HTTPS access is required, you need to bind an HTTPS certificate. For more information, see Configure Domain Name Resolution and Bind Certificate.

Inventory Overview

Last updated：2025-12-08 16:19:03

What Is Inventory

The inventory feature helps users manage objects in a bucket. It operates periodically as an alternative to the COS synchronous List API. Cloud Object Storage (COS) can scan specified objects or objects with the same prefix in user buckets daily, weekly, or monthly based on user-configured inventory tasks, and generate an inventory report stored as a CSV file in the specified bucket. It also supports triggering a one-time inventory task to immediately produce a report. The file lists stored objects and their corresponding metadata, recording required object property information according to user configuration.
You can use the inventory feature for various purposes, including but not limited to:
Reviewing and reporting the replication and encryption status of objects.
Simplifying and accelerating business workflows and big data jobs.
Note:
You can configure multiple inventory jobs in one bucket. Such jobs do not directly read the object content during their execution; instead, they only scan the attribute information such as object metadata.
Currently not supported in financial cloud regions.
To set up an inventory through a sub-account, please ensure the sub-account has been granted cos:PutBucketInventory and cos:PutObject operation permissions by the root account.

Inventory Parameters

After you configure an inventory job for a bucket, COS regularly scans specified objects in the bucket based on your configuration, and exports an inventory report in .csv format. Currently, the inventory report can record the following information:
Inventory Information
Description
AppID
Account ID.
Bucket
Name of the bucket where the inventory job is performed.
fileFormat
File format.
listObjectCount
Number of objects to list, which will be used for billing. For more information, see the inventory feature fees in Management Fees.
listStorageSize
Size of listed objects.
filterObjectCount
Number of filtered objects.
filterStorageSize
Size of filtered objects.
Key
Name of an object file in a bucket. When the CSV file format is used, the key name is URL-encoded and must be decoded before you can use it.
VersionId
Version ID of an object. After versioning is enabled for a bucket, COS specifies a version ID for the object added to the bucket. If the inventory is for the current version only, this field is not included.
IsLatest
Set to True if the latest object version is used. This field is not included if the inventory is only for the current version of objects.
IsDeleteMarker
Set to True if the object is a delete marker. This field is not included if the inventory is only for the current version of objects.
Size
Object size (measured in bytes)
LastModifiedDate
The last modified date of the object.
ETag
Hash value of an object. It displays only modification to the content of an object, rather than to its metadata. The ETag may be or may not be the MD5 checksum of the data of the object, and this depends on how the object is created and encrypted.
StorageClass
Storage class of the object. For more information, see Storage Class.
IsMultipartUploaded
Set to True if the object is uploaded in multiple parts. For more information, see Multipart Upload.
Replicationstatus
Used for marking the status of source files and replica files in object replication. Source file marks: PENDING (awaiting replication), COMPLETED (replication completed), FAILED (replication failed); replica file marks: REPLICA (replication completed, replica file generated). For more information, please refer to Replication Behavior Description.
Tag
Object tag.
Crc64
CRC64 checksum of the object.
x-cos-meta-*
Custom metadata of the object, such as x-cos-meta-testheader. The inventory outputs the corresponding object metadata. If the object does not contain the metadata, it is empty.

Periodic Inventory and One-Time List

The inventory supports two types: periodic inventory and one-time list.
Periodic inventory: Allows users to set inventory rules to generate inventory reports by month, week, or day. Among them, monthly requires specifying the output date (supports setting the 1st–28th), while weekly continues to generate reports on the configured day of the week.
One-time inventory: Initiate a one-time inventory task to generate a report. Currently, COS does not support record query for one-time inventory tasks, so save the task initiation Rule id, inventory storage path, etc. for future reference when searching the inventory report.

How to Configure Inventory

Before configuring an inventory, you need to understand two concepts:
Source bucket: The bucket for which you want to enable inventory, containing:
Objects listed in the inventory.
Configuration of the inventory.
Destination bucket: The bucket where to store the inventory report, containing:
An inventory list file.
Manifest files that describe the location of the inventory list file.
An inventory can be configured as follows:

Specify the information of the objects to be analyzed in the source bucket

To inform COS of object information to be analyzed, you need to configure the following information in the source bucket for the inventory:
Select object versions: List all the versions of the object or list only the current version. If you select to list all the versions of the object, COS adds information about the object in all historical versions to the inventory report. If you select to list only the current version, COS records only the object in the latest version.
Configure the attributes of the objects to be analyzed: You need to tell COS which information in the object attributes needs to be included in the inventory report. Currently, supported object attributes include account ID, source bucket name, object file name, object version ID, whether the object is of the latest version, whether the object is a deletion flag, object size, object's last modified date, ETag, object's storage class, cross-region replication tag, and whether the object is a part in multipart upload.

Configure the storage information for the inventory report

You need to tell COS how often to export the inventory report, which bucket to store the report in, and whether the report should be encrypted. The configuration information is as follows:
Select an export frequency: Daily or weekly. COS will execute the inventory feature at the specified frequency.
Select an encryption mode: No encryption or SSE-COS. If SSE-COS is selected, COS encrypts the generated inventory report.
Configure the output location of the inventory: You need to specify the bucket where to store the inventory report.
Note:
The destination bucket must be in the same region as the source bucket. They can be the same bucket.

How to Use

Configuring Inventory via the Console

Adding a Scheduled List

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List.
3. Click to enter the bucket (source bucket) for which you want to enable the inventory feature.
4. In the left navigation bar, select basic configuration > inventory setting, then click Add list.
5. On the configuration page, you can configure the following:
5.1 Basic information
Rule Status: Select whether to enable the current inventory rule. You can select to enable/disable the inventory.
Inventory Name: Type the name of your inventory.
Target bucket: Refers to the bucket where the inventory is stored. The initial default destination bucket is the source bucket. The destination bucket must be in the same region as the source bucket.
Inventory report prefix (optional): Input a prefix of your choice for the target storage bucket, which can be used to group inventory files in a common location. The initial value is the default value.
Inventory report generation path: Refers to the path where the inventory report is generated.
5.2 Filtering Condition
File range: Select the scope of the inventory objects. You can select the entire bucket or some files with the same prefix. For example, to select all files in the doc path under the folder path in the bucket, specify the file prefix as folder/doc.
Object Version: Select whether the inventory will include all versions of the object or only the current version. If this is not selected, only the current version will be kept by default.
Object tag (optional): Filter objects with the same tag into the inventory. If left blank, no filtering is performed by default.
Filtering Time (optional): List objects whose modification time is after a certain time point or within a certain time to the inventory through filtering. If this is not set, no filtering will be performed by default.
Storage Type Filter (optional): List objects with the specified storage type (if any) into the inventory. For an introduction to storage types, please refer to Storage Type Overview. If not set, no filtering is performed by default.
Custom output header (optional): If you need to include custom headers of the object in the inventory report, you can fill in the custom headers to be output. Only x-cos-meta-* headers are supported. If not filled in, the custom header will not be output by default.
Inventory Information: Select the corresponding information of the object to be included in the inventory report. Options include: size, storage type, ETag, bucket replication status, multipart upload status, latest update date, Tag, CRC64. If not selected, all are selected by default.
Note:
ETag (Entity Tag) is a hash of an object. ETag reflects only changes to the content of an object but not changes to its metadata. The ETag is not necessarily an MD5 digest of the object's data, depending on how the object was created and encrypted.
The generated inventory content by default includes the field contents of AppID, Bucket, Key, and LastModifiedTime.
If versioning is enabled for the bucket, the generated inventory contents increase with VersionId, IsLatest, and IsDeleteMarker fields.
5.3 Output Format
Output format: The output inventory file format is a CSV file compressed using GZIP.
Generation Cycle: Select whether to generate the inventory daily/weekly/monthly. The default is to generate the inventory daily. The specific time to generate the inventory depends on the amount of data. The description of the configuration is shown below:
Daily: For example, if the list is added at 3 PM on the same day, the list file is typically generated and delivered to the target bucket by 6 AM the next day (estimated time). If the data volume is large, the generation time may be delayed. It is recommended to choose weekly or monthly list generation.
Weekly: The inventory is generated every 7 days by default. For example, if the inventory is enabled on September 1, the first inventory will be generated on September 2, then on September 9, 16, 23... at 7-day intervals.
Per month: You can specify any day from the 1st to the 28th to generate the inventory. The generation logic is as follows:
If the current day is the specified date when creating the inventory, it will be generated on the specified date of the current month.
For example, if you enable inventory on September 8 and designate the 25th of each month for generation, the first inventory will be generated on September 25, followed by October 25, November 25, December 25, and so on, with periodic inventory generation by month.
If the current day is on or after the specified date when creating the inventory, it will be generated on the specified date of the next month.
For example, if you enable inventory on September 8 and designate the 3rd of each month for generation, the first inventory will be generated on October 3, then on November 3, December 3, and so on, with periodic inventory generation by month.
Inventory Encryption: Select whether the inventory needs server-side encryption. The options are:
No encryption: The inventory will not be encrypted. This is the default option.
SSE-COS: Encrypt reports using server encryption with COS-managed keys. For more information, see SSE-COS Encryption in the COS Developer Guide.
Access authorization: Disabled by default. After enabling access authorization, you can proceed to the next step.
5.4 Information Confirmation
5.5 Confirm the inventory configuration information for the bucket. If modification is needed, click Back.
6. After confirming that the configuration information is correct, click Confirm. COS will then automatically generate inventory files based on the set period and deliver them to the target bucket you set.
Note:
For details on the format and content of the inventory report generated by the inventory feature, see Feature Overview of Inventory.
7. After successful creation, the inventory will be delivered to the specified path, which is the inventory report generation path in 5.1. For more information, see inventory report storage path.

Generating an Instant List

Note:
If you want to immediately execute created inventory rules, select instant inventory generation.
1. Log in to the COS Console.
2. In the left sidebar, click Bucket List, then click to enter the created bucket (source bucket) with inventory rules.
3. Click on the left basic configuration > inventory setting, find the corresponding inventory rule, then click on the right more > immediately execute to quickly generate a list.

4. After successful creation, the instant inventory will be delivered to the specified path, which is the current scheduled inventory rule's configured inventory report generation path.

Previewing Inventory

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List, then click to enter the created bucket (source bucket) with inventory rules.
3. Click on the left basic configuration > inventory setting, find the corresponding inventory rule, then click more > preview inventory to quickly view the inventory result.

4. The preview page shows all inventories generated under the current rule, including scheduled inventory and instant inventory. You can go to the directory and view inventory results with one-click, and support filtering by date and inventory type.
Note:
After triggering the manifest task, the inventory result will update gradually. The result will be fully generated only after the manifest.json file is generated under the directory.

5. After selecting the date, in the directory name column, click a directory to quickly redirect to it. You can view the generated inventory contents. For more details, please refer to inventory report storage path.

6. To directly view the inventory result, click View Inventory Result in the operation column, and you will see the total number of objects and sizes under your configuration criteria. For more detailed inventory information, click Download to decompress and view.

7. The field names corresponding to the downloaded CSV file can be viewed in the "fileSchema" of the manifest.json file. For more field meanings and descriptions, please refer to Inventory Parameters.


Configuring an Inventory Through APIs

To enable the inventory feature for a specified bucket using APIs, follow the steps below:
1. Create a COS role.
2. Bind permissions to the COS role.
3. Enable inventory.

1. Create a COS Role

Create a COS role, for detailed API information, refer to CreateRole. The roleName must be: COS_QcsRole. The policyDocument is:
{
    "version": "2.0",
    "statement": [{
        "action": "name/sts:AssumeRole",
        "effect": "allow",
        "principal": {
            "service": "cos.cloud.tencent.com"
        }
    }]
}

2. COS Role Binding Permissions

Bind role permissions, for more information on the specific interface, see AttachRolePolicy. Here, policyName is: QcloudCOSFullAccess, and roleName is the COS_QcsRole from step 1, or you can use the roleID returned when creating the roleName.

3. Enable the Inventory Feature

The destination bucket for storing inventory files must be in the same region as the source bucket. Call the API to enable inventory. The API information is as follows:
Periodic Inventory
Configure periodic inventory: PUT Bucket inventory
View periodic inventory rules: GET Bucket Inventory
Delete periodic inventory rules: DELETE Bucket Inventory
List periodic inventory rules: LIST Bucket Inventory
Trigger a one-time inventory task: POST Bucket Inventory

Inventory Report Storage Path

The manifest reports and associated manifest files will be published in the destination bucket, with the manifest reports being published at the following path:
destination-prefix/appid/source-bucket/config-ID/
The manifest files will be published under the following path in the destination bucket:
destination-prefix/appid/source-bucket/config-ID/YYYYMMDD/manifest.json
destination-prefix/appid/source-bucket/config-ID/YYYYMMDD/manifest.checksum
Note:
After initiating an inventory task, the results will be progressively updated. The inventory results will only be fully generated once the manifest.json file under the inventory directory is created.
The meanings of the paths are as follows:
destination-prefix: This is the "destination prefix" set by the user when configuring the inventory. It can be used to group all inventory reports in a common location within the target bucket.
source-bucket: This is the name of the source bucket corresponding to the inventory report. This folder is added to avoid conflicts that may occur when multiple source buckets send their inventory reports to the same destination bucket.
config-ID: This is the "inventory name" set when you configure the inventory. If multiple inventory reports are configured in the same source bucket and delivered to the same destination bucket, config-ID can be used to distinguish among different reports.
YYYYMMDD: Timestamp, including the time and date when the bucket scan is started when the inventory report is generated.
manifest.json: This is the manifest file.
manifest.checksum: This is the MD5 of the content of the manifest.json file.
There are two manifest files: manifest.json and manifest.checksum.
Note:
Below describes the Manifest files:
Both manifest.json and manifest.checksum are Manifest files. The manifest.json file describes the location of the inventory report, while manifest.checksum serves as the MD5 of the manifest.json file content. Each time a new inventory report is delivered, a new set of Manifest files is included.
Each manifest contained in manifest.json provides metadata and basic information about the inventory, including:
Source bucket name.
Destination bucket name.
Inventory version.
Timestamp, including the time and date when the bucket scan is started when the inventory report is generated.
Format and architecture of the inventory file.
Object key, size, and MD5 checksum of the inventory report in the destination bucket.
The following example shows a manifest in the manifest.json file for CSV-formatted inventory:
{
 "sourceAppid": "1250000000",
 "sourceBucket": "example-source-bucket",
 "destinationAppid": "1250000000",
 "destinationBucket": "example-inventory-destination-bucket",
 "fileFormat": "CSV",
 "listObjectCount": "13",
 "listStorageSize": "7212835",
 "filterObjectCount": "13",
 "filterStorageSize": "7212835",
 "fileSchema": "Appid, Bucket, Key, Size, LastModifiedDate, ETag, StorageClass, IsMultipartUploaded, ReplicationStatus",
 "files": [
  {
   "key": "cos_bucket_inventory/1250000000/examplebucket/inventory01/04d73d9debc73d9f0bf85af461abde6c.csv.gz",
   "size": "502",
   "md5Checksum": "7d40288a09c25b302ad6cb5fced54f35"
  }
 ]
}

Inventory Consistency

All of your objects might not appear in each inventory list. The inventory list provides eventual consistency for PUTs of both new objects and overwrites, and DELETEs. Therefore, the inventory report possibly does not include the latest added or deleted object. For example, if the user uploads or deletes an object when COS is executing an inventory job configured by the user, the results of the upload or deletion operations may not be reflected in the inventory report.
If you want to verify the object status before the execution, you are advised to call the HEAD Object API to extract the object metadata, or check the object attributes in the COS console.

Bucket Tag Overview

Last updated：2025-09-24 17:45:28

Overview

A bucket tag is a key-value pair (key = value), which is comprised of a tag key and a tag value that are connected by an equal sign ("="), for example: group = IT. It can be used to manage buckets in groups. You can set, query, and delete the tag for a specified bucket.

Specifications and Limits

Limits on Tag Keys

A tag key starting with "qcs:" or "project" is a default tag key and cannot be created.
A tag key can only contain characters encoded in UTF-8, spaces, numbers, or special characters including + - = . _ : / @.
A tag key should be a combination of 1-127 characters encoded in UTF-8.
Tag keys are case sensitive.

Limits on Tag Values

A tag value can only contain characters encoded in UTF-8, spaces, numbers, or special characters including + - = . _ : / @.
A tag value should be a combination of 1-255 characters encoded in UTF-8.
Tag values are case sensitive.

Limits on the Number of Tags

Bucket dimension: A resource has at most 50 different bucket tags.
Tag dimension:
A user has at most 1,000 different keys.
A key has at most 1,000 values.
Multiple same keys are not allowed in the same bucket.

Usage

You can set bucket tags via the console, APIs/SDKs, or tools.

Using the COS Console

Adding a Tag When Creating a Bucket

You can add a bucket tag when creating a bucket, as shown in the following figure:


Adding a Tag to an Existing Bucket

If you did not add tags when you created the bucket, you can follow the steps below to add tags to the bucket.
1. On the bucket list page, find the bucket you need to add a tag to, click its name, and go to the bucket configuration page.
2. Click on the left Basic Configurations, scroll down the page to find the Tagging configuration item, and click Add Tags.

The configuration items are described as follows:
Tag keys: case-sensitive, support characters such as Chinese, a-z, A-Z, 0-9, +, -, _, =, /, ., :, @.
Tag values: case-sensitive, support characters such as Chinese, a-z, A-Z, 0-9, +, -, _, =, /, ., :, @.
Note:
A bucket can have a maximum of 50 tags, and tag keys cannot be repeated.
3. After entering the tag key and tag value, click Save to complete adding a bucket tag.

Using REST API/SDK

You can use the following APIs to manage bucket tags directly:
PUT Bucket tagging
GET Bucket tagging
DELETE Bucket tagging
You can manage bucket tags directly using the following SDK:
Android, C, C++, .NET(C#), Go, iOS, Java, JavaScript, Node.js, PHP, Python, Mini Program SDK.

Tools

Manage bucket tags directly using the following tools:
COSBrowser
COSCLI



Object Tag Overview

Last updated：2025-09-24 17:45:28

Overview

Object tagging is designed to help you group and manage objects in your bucket by adding a key-value pair as an object tag. An object tag consists of a tagKey, a =, and a tagValue, such as group = IT. You can set, query, and delete tags on the specified object.
Note:
Object tagging is a paid feature. For detailed pricing, see Pricing | Cloud Object Storage.

Specifications and Restrictions

Tag key restrictions

Supports characters such as Chinese, a-z, A-Z, 0-9, +, -, _, =, /, ., :, @, and distinguishes between uppercase and lowercase letters in English.
A tag key can contain 1–127 UTF-8 characters.
Tag keys are case-sensitive.

Tag value restrictions

Supports characters such as Chinese, a-z, A-Z, 0-9, +, -, _, =, /, ., :, @, and distinguishes between uppercase and lowercase letters in English.
A tag value can contain 1–255 UTF-8 characters.
Tag values are case-sensitive.

Tag quantity restrictions

Object dimension: Up to ten unique tags per object.
Tag dimension: Unlimited.

Usage

Using the COS Console

Adding tag during object upload

1. Add object tags when uploading object, as shown below:

2. After a successful upload, object tags can be added. You can go to the bucket's file list page, find the object with added tags, click More > Add tag to view, edit or delete the added tags.

Adding tag uploaded object

If you did not add tags when you uploaded the object, you can add tags to the object by following the steps below.
1. See view object information and go to the file list page.
2. On the file list page, find the object to be operated, click More > Add tag.
3. In the pop-up window, click Add tag, manually input the tag key and value, then save.
If you need to modify or delete a tag, herein in a window, click Edit or Delete.

Using REST API/SDK

You can manage object tags via the following APIs:
PUT Object tagging
GET Object tagging
DELETE Object tagging
You can call the management object tag method in the SDK. For details, see the following language SDK documents:
Android, .NET, Go, iOS, Java, JavaScript, Node.js, PHP, Python, mini program.

Tools

You can manage object tags directly using the COSCLI tool.

FAQs

How to Use Object Tags

After setting up object tags, you can quickly set lifecycle rules for objects with the same tags. For details, see Setting the Bucket Lifecycle.




Event Notifications

Last updated：2024-03-25 15:28:24

Overview

If any change is made to Cloud Object Storage (COS) resources (such as new files uploaded or files deleted), you will receive prompt notification. Event notification can be used together with Serverless Cloud Function (SCF) to meet the needs of more application scenarios:
Interaction with other Tencent Cloud services: For example, purge CDN cache or update the database when a new file is uploaded to COS.
System integration: Your own service APIs can be called when files in COS are created, deleted, or overwritten. In user-generated content (UGC) scenarios, with the event notification feature, the client side will be able to interact with the server side.
Data processing: Files in COS can be automatically processed, such as automatic decompression and AI recognition.
COS event notification

COS event notification has the following features:
Async processing: sending notifications does not affect normal COS operations.
Notification targets: notifications can only be sent to SCF functions in the same region.
Currently, an SCF function can be triggered by the following COS events:
Event Type
Description
cos: ObjectCreated:*
All upload events mentioned below
cos: ObjectCreated:Put
An object is created using the Put Object API
cos: ObjectCreated:Post
An object is created using the Post Object API
cos: ObjectCreated:Copy
An object is created using the Put Object - Copy API
cos: ObjectCreated:CompleteMultipartUpload
An object is created using the Complete Multipart Upload API
cos:ObjectCreated:Origin
Image origin-pull occurs
cos:ObjectCreated:Replication
An object is created using cross-region replication
cos: ObjectRemove:*
All deletion events mentioned below
cos: ObjectRemove:Delete
An object is deleted from an unversioned bucket using the Delete Object API, or an object with a specified version is deleted by specifying versionid
cos: ObjectRemove:DeleteMarkerCreated
An object is deleted from a versioning-enabled or versioning-suspended bucket using the Delete Object API
cos:ObjectRestore:Post
An ARCHIVE restoration job is created
cos:ObjectRestore:Completed
An ARCHIVE restoration job is completed


How to Use COS Event Notification

Please follow the steps below:
1. Create an SCF function
You can create a function in the SCF console or using the CLI. When creating the function, you need to select the runtime environment based on the language you will use to write the function and submit the function code either by editing online or uploading a local code package.
You can also use a preconfigured SCF template to simplify the creation. For more information, see Create a Function. The way to write the function varies by programming language. For more information, see the SCF documentation.
2. Test the function Once a function is created, you can test it with a test template which can simulate COS events and trigger the function. For more information, see Testing a Function.
3. Add a trigger After you finish the testing, you can bind the SCF function with a bucket by creating a COS trigger in the console or using the command line. For more information, see Creating a Trigger.
4. Check if the function works After completing the steps above, you can make changes to the bucket in COS to see if everything works. For example, you can upload or delete files via the console or COSBrowser. Then, go to the SCF console, choose Function Service, click the name of the target function, and click Log Query to check if everything works properly.
For more information on SCF COS triggers, see COS Trigger.

Data Extraction

SELECT Overview

Last updated：2025-10-29 16:10:51

The COS Select feature uses Structured Query Language (SQL) statements to filter the objects stored in COS so as to extract specific objects and get desired data. With COS Select, you can reduce the amount of data transferred by COS for lower costs and latency during data extraction.
The COS Select feature currently allows you to extract objects stored in CSV, JSON, and Parquet formats and compressed by gzip or bzip2 (for CSV and JSON objects only). In addition, you can save extraction results in CSV and JSON formats and specify how to separate the result records.
You can pass in a SQL expression to COS in your request. COS Select currently only supports certain SQL expressions. For more information, see SQL Functions.
You can use the COS console, API, SDK, or COSCMD to perform SQL queries. Note that certain limits apply to file extraction in the COS console: up to 128 MB of files can be extracted, and up to 40 MB of data can be returned. To extract more data, you need to use other methods.
Note:
For more information on data types supported by COS Select and current reserved fields, see Data Types and Reserved Fields.
Currently, the extraction feature is only supported in the Chinese mainland public cloud regions and Virginia, and is not available in other regions.

Restrictions

The following restrictions apply to COS Select:
You must have the cos:GetObject permission to the queried object. A root account has this permission by default.
Only objects in the STANDARD storage class can be extracted.
The maximum length of a SQL expression is 256 KB.
The maximum length of a single record in the extraction result is 1 MB.
SQL clauses currently supported by COS Select include:
SELECT statement
FROM clause
WHERE clause
LIMIT clause
Note:
For more information on SQL clauses, see SELECT Command.
Functions currently supported by COS Select include:
Aggregate functions, such as AVG, COUNT, MAX, MIN, and SUM.
Condition functions, such as COALESCE and NULLIF.
Conversion functions, such as CAST.
Date functions, such as DATE_ADD, DATE_DIFF, EXTRACT, TO_STRING, TO_TIMESTAMP, and UTCNOW.
String functions, such as CHAR_LENGTH, CHARACTER_LENGTH, LOWER, SUBSTRING, TRIM, and UPPER.
Note:
For more information on SQL functions, see SQL Functions.
COS Select currently supports the following operators:
Logical operators: AND, NOT, OR
Comparison operators: <, >, <=, >=, =, <>, !=, BETWEEN, IN
Pattern matching operators: LIKE
Mathematical operators: +, -, *, %
Note:
For more information on operators, see Operators.

Initiating Extraction Request

You can initiate an extraction request using the console, API, or SDK:
If you use the console, see Data Extraction.
If you use the API, see SELECT Object Content.
If you use the SDK, go to SDK Overview and select the required SDK API.

FAQs

If a problem occurs when you perform a query, COS Select will return an error code and error message. For the list of error codes and descriptions, see Special Error Codes.

SELECT Command

Last updated：2024-03-25 15:28:24

Overview

The COS Select feature only supports the SELECT command to extract the required data and reduce the amount of data transferred, which helps lower the costs and request delay. The following are the standard clauses supported by SELECT queries:
SELECT statement
WHERE clause
LIMIT clause
COS Select currently does not support clause queries or joins.

SELECT Statement

The SELECT statement can extract the data you want to see from a COS object. You can query the data at different dimensions such as column name, function, and expression, and the query result will be returned as a list. The format of SELECT statement call is as follows:
SELECT *
SELECT projection [ AS column_alias | column_alias ] [, ...]
The first SELECT statement is marked with * (asterisk) and will return all the columns in the COS object. The second one uses a user-defined output scalar expression, and projection creates a list of outputs with custom names for each column.

WHERE Clause

The WHERE clause uses the following syntax:
WHERE condition
The WHERE clause filters data by condition. condition is an expression that returns a Boolean result, and only rows with a return value of TRUE will be output in the result.

LIMIT Clause

The LIMIT clause uses the following syntax:
LIMIT number
The LIMIT clause sets a limit on the number of records to be returned per query, which can be specified using the number parameter.

Access Attributes

The SELECT and WHERE clauses can select the fields to be queried in any of the following ways, depending on whether the file format is CSV or JSON.

CSV

Column number: You can use _N to specify the data in column N for query. For any CSV files, the column number increases from 1. For example, the first column is numbered _1, and the second column is numbered _2. In the SELECT and WHERE clauses, it is valid to specify the column to be queried using _N or alias._N.
Column header: If the CSV file to be queried contains column headers, the SELECT and WHERE clauses can use the headers to specify the columns to be queried, which can be specified using alias.column_name or column_name in the SELECT and WHERE clauses in an SQL statement.

JSON

Document: You can access a JSON document using alias.name. A nested array can be accessed in a way such as alias.name1.name2.name3.
List: You can access the elements in a list using an index, which is numbered from 0 and uses the [] operator. For example, you can access the second element in a JSON list using alias[1]. If you need to access a nested array, you can also do so in a way such as alias.name1.name2[1].name3.
Samples The following is the sample data in the samples:
{
    "name": "Leon",
    "org": "Tencent",
    "projects":
        [
         {"project_name":"project1", "completed":true},
         {"project_name":"project2", "completed":false}
        ]
}
Sample 1. The following is the SQL statement used to query name in the sample data and the query result:
Select s.name from COSObject s
{"name":"Leon"}
Sample 2. The following is the SQL statement used to query project_name in the sample data and the query result:
Select s.projects[0].project_name from COSObject s
{"project_name":"project1"}

Case Sensitivity of Headers and Attribute Names

You can use double quotation marks to indicate whether headers in a CSV file and attribute names in a JSON file are case-sensitive. If no double quotation marks are added, the headers/attribute names are case-insensitive. If this is not explicitly specified, COS Select may throw an exception.
Sample 1. Query objects with a header/attribute name containing "NAME". Because the following sample SQL query does not contain double quotation marks, the query is case-insensitive. As this header is present in the table, a value will be successfully returned eventually.
SELECT s.name from COSObject s
Because the following sample SQL query contains double quotation marks, the query is case-sensitive. As the table does not contain this header, the SQLParsingError 400 error will be eventually returned.
SELECT s."name" from COSObject s
Sample 2. Query objects with a header/attribute name containing "NAME" and "name". Because the following sample SQL query does not contain double quotation marks, the query is case-insensitive. As the table contains two headers "NAME" and "name", the query command is ambiguous, and the AmbiguousFieldName exception will be thrown.
SELECT s.name from COSObject s
Because the following sample SQL query contains double quotation marks, the query is case-sensitive. As the table contains the header "NAME", the query result will be successfully returned.
SELECT s."NAME" from COSObject s

Using Reserved Fields as User-defined Fields

The SQL expressions of COS Select have certain reserved fields such as function name, data type, and operator. Sometimes you probably use these reserved fields as column headers in a CSV file or attribute names in a JSON, which may cause conflicts with reserved fields. In this case, you can use double quotation marks to indicate that you are using a custom field; otherwise, COS will return 400 parse error.
For the complete list of reserved fields, see Reserved Words.
Sample: The header/attribute name in the object to be queried contains a reserved field "CAST". The following sample SQL query uses double quotation marks to indicate that CAST is a user-defined field, so the query result will be successfully returned.
SELECT s."CAST" from COSObject s
The following sample SQL query does not use double quotation marks to indicate that CAST is a user-defined field, so COS will treat it as a reserved field and return 400 parse error.
SELECT s.CAST from COSObject s

Scalar Expressions

In the SELECT statement and the WHERE clause, you can use SQL scalar expressions (expressions that return a scalar). Currently, COS Select supports the following forms:
literal: SQL text.
column_reference: column_name or alias.column_name.
unary_opexpression: SQL unary operator.
expressionbinary_opexpression: SQL binary operator.
func_name: Name of the called scalar function.
expression [ NOT ] BETWEEN expression AND expression
expression LIKE expression [ ESCAPE expression ]

SQL Functions

Last updated：2024-03-25 15:28:24

Aggregate Functions

COS Select supports the following aggregate functions:
Function Name
Parameter Type
Return Type
AVG(expression)
INT, FLOAT, and DECIMAL
DECIMAL will be returned if the input parameter is of integer type, and FLOAT if float type. The same type as the input parameter will be returned in all other cases.
COUNT
-
INT
MAX(expression)
INT and DECIMAL
The return value type is the same as that of the input parameter
MIN(expression)
INT and DECIMAL
The return value type is the same as that of the input parameter
SUM(expression)
INT, FLOAT, DOUBLE, and DECIMAL
INT will be returned if the input parameter is of integer type, and FLOAT if float type. The same type as the input parameter will be returned in all other cases.

Condition Functions

COS Select supports the following condition functions:

COALESCE

The COALESCE function determines the input parameters in sequence and returns the first non-null parameter value. If the input parameters do not include a non-null parameter, the function will return a null value.

Syntax

COALESCE ( expression, expression, ... )
Note:
Values, arrays, or nested functions of INT, String, and Float types can be passed in for the expression parameter.

Sample

COALESCE(1)                -- 1
COALESCE(1, null)          -- 1
COALESCE(null, null, 1)    -- 1
COALESCE(missing, 1)       -- 1
COALESCE(null, 'string')   -- 'string'
COALESCE(null)             -- null
COALESCE(null, null)       -- null
COALESCE(missing)          -- null
COALESCE(missing, missing) -- null

NULLIF

The NULLIF function determines the difference between two parameters passed in. If the two parameters have the same value, NULL will be returned; otherwise, the value of the first parameter passed in will be returned.

Syntax

NULLIF ( expression1, expression2 )
Note:
Values, arrays, or nested functions of INT, String, and Float types can be passed in for the expression parameter.

Sample

NULLIF(1, 2)             -- 1
NULLIF(1, '1')           -- 1
NULLIF(1, NULL)          -- 1
NULLIF(1, 1)             -- null
NULLIF(1.0, 1)           -- null
NULLIF(missing, null)    -- null
NULLIF(missing, missing) -- null
NULLIF([1], [1])         -- null
NULLIF(NULL, 1)          -- null
NULLIF(null, null)       -- null

Conversion Functions

COS Select supports the following conversion functions:

CAST

The CAST function converts one instance to another instance. The instance can be either a value or a function that can be calculated to a certain value.

Syntax

CAST ( expression AS data_type )
Note:
The expression parameter can be a value, an array, an operator, or an SQL function that can be calculated to a certain value.
The data_type parameter is the data type after conversion, such as INT. For the data types currently supported by COS Select, see Data Types.

Sample

CAST('2007-04-05T14:30Z' AS TIMESTAMP)
CAST(0.456 AS FLOAT)

Date Functions

COS Select supports the following date functions:

DATE_ADD

The DATE_ADD function adds a specified time interval to a part (year, month, day, hour, minute, or second) of a specific timestamp and returns a new timestamp.

Syntax

DATE_ADD( date_part, quantity, timestamp )
Note:
The date_part parameter specifies the part of the timestamp to be modified, which can be year, month, day, hour, minute, or second.
The quantity parameter represents the value to be added, which must be a positive integer.
The timestamp parameter represents the timestamp to be modified.

Sample

DATE_ADD(year, 5, `2010-01-01T`)                -- 2015-01-01
DATE_ADD(month, 1, `2010T`)                     -- 2010-02T 
DATE_ADD(month, 13, `2010T`)                    -- 2011-02T
DATE_ADD(hour, 1, `2017T`)                      -- 2017-01-01T01:00-00:00
DATE_ADD(hour, 1, `2017-01-02T03:04Z`)          -- 2017-01-02T04:04Z
DATE_ADD(minute, 1, `2017-01-02T03:04:05.006Z`) -- 2017-01-02T03:05:05.006Z
DATE_ADD(second, 1, `2017-01-02T03:04:05.006Z`) -- 2017-01-02T03:04:06.006Z

DATE_DIFF

The DATE_DIFF function compares two valid timestamps and returns the difference between them, which can be displayed in the specified unit of time. If the date_part value of timestamp1 is greater than that of timestamp2, a positive number will be returned; otherwise, a negative number will be returned.

Syntax

DATE_DIFF( date_part, timestamp1, timestamp2 )
Note:
The date_part parameter specifies the unit of time which the two timestamps are compared in and can be year, month, day, hour, minute, or second.
The timestamp1 parameter is the first input timestamp.
The timestamp2 parameter is the second input timestamp.

Sample

DATE_DIFF(year, `2010-01-01T`, `2011-01-01T`)            -- 1
DATE_DIFF(year, `2010T`, `2010-05T`)                     -- 4 
DATE_DIFF(month, `2010T`, `2011T`)                       -- 12
DATE_DIFF(month, `2011T`, `2010T`)                       -- -12
DATE_DIFF(day, `2010-01-01T23:00T`, `2010-01-02T01:00T`) -- 0

EXTRACT

The EXTRACT function extracts a value in the specified unit of time from a given timestamp.

Syntax

EXTRACT( date_part FROM timestamp )
Note:
The parameter date_part specifies the unit of time to be extracted, which can be year, month, day, hour, minute, or second.
The timestamp parameter represents the input timestamp.

Sample

EXTRACT(YEAR FROM `2010-01-01T`)                           -- 2010
EXTRACT(MONTH FROM `2010T`)                                -- 1 
EXTRACT(MONTH FROM `2010-10T`)                             -- 10
EXTRACT(HOUR FROM `2017-01-02T03:04:05+07:08`)             -- 3
EXTRACT(MINUTE FROM `2017-01-02T03:04:05+07:08`)           -- 4
EXTRACT(TIMEZONE_HOUR FROM `2017-01-02T03:04:05+07:08`)    -- 7
EXTRACT(TIMEZONE_MINUTE FROM `2017-01-02T03:04:05+07:08`)  -- 8

TO_STRING

The TO_STRING function converts a timestamp to a string of time in the specified format.

Syntax

TO_STRING ( timestamp time_format_pattern )
Note:
The timestamp parameter specifies the timestamp to be converted.
The time_format_pattern parameter specifies the time format.
Format
Description
Sample
yy
2-digit year
98
y
4-digit year
1998
yyyy
Year expressed by 4 digits. If there are less than 4 digits, 0 will be automatically added
0199
M
Month
1
MM
Month expressed by 2 digits. If there are less than 2 digits, 0 will be automatically added
01
MMM
English abbreviation of a month
Jan
MMMM
Full English name of a month
January
MMMMM
Initial of a month
J (not applicable to the to_timestamp function)
d
Day (1-31) in a month
1
dd
Day expressed by 2 digits (1-31)
01
a
Symbol for morning or afternoon (AM/PM)
AM
h
Hour in 12-hour time
1
hh
Hour expressed by 2 digits in 12-hour time
01
H
Hour in 24-hour time
1
HH
Hour expressed by 2 digits in 24-hour time
01
m
Minute (00-59)
1
mm
Minute expressed by 2 digits in 24-hour time
01
s
Second (00-59)
1
ss
Second expressed by 2 digits in 24-hour time
01
S
Decimal part of the second (accuracy: 0.1; value range: 0.0 - 0.9)
0
SS
Decimal part of the second (accuracy: 0.01; value range: 0.00 - 0.99)
6
SSS
Decimal part of the second (accuracy: 0.001; value range: 0.000 - 0.999)
60
...
...
...
SSSSSSSSS
Decimal part of the second (accuracy: 0.000000001; value range: 0.000000000 - 0.999999999)
60000000
n
Nanosecond
60000000
X
Hour-level offset. If the offset is 0, then this will be "Z"
+01 or Z
XX or XXXX
Hour- or minute-level offset. If the offset is 0, then this will be "Z"
+0100 or Z
xxx or xxxxx
Hour- or minute-level offset. If the offset is 0, then this will be "Z"
+01:00 or Z
x
Hour-level offset
1
xx or xxxx
Hour- or minute-level offset
0100
xxx or xxxxx
Hour- or minute-level offset
01:00

Sample

TO_STRING(`1998-07-20T20:18Z`,  'MMMM d, y')                    -- "July 20, 1998"
TO_STRING(`1998-07-20T20:18Z`, 'MMM d, yyyy')                   -- "Jul 20, 1998"
TO_STRING(`1998-07-20T20:18Z`, 'M-d-yy')                        -- "7-20-69"
TO_STRING(`1998-07-20T20:18Z`, 'MM-d-y')                        -- "07-20-1998"
TO_STRING(`1998-07-20T20:18Z`, 'MMMM d, y h:m a')               -- "July 20, 1998 8:18 PM"
TO_STRING(`1998-07-20T20:18Z`, 'y-MM-dd''T''H:m:ssX')           -- "1998-07-20T20:18:00Z"
TO_STRING(`1998-07-20T20:18+08:00Z`, 'y-MM-dd''T''H:m:ssX')     -- "1998-07-20T20:18:00Z"
TO_STRING(`1998-07-20T20:18+08:00`, 'y-MM-dd''T''H:m:ssXXXX')   -- "1998-07-20T20:18:00+0800"
TO_STRING(`1998-07-20T20:18+08:00`, 'y-MM-dd''T''H:m:ssXXXXX')  -- "1998-07-20T20:18:00+08:00"

TO_TIMESTAMP

The TO_TIMESTAMP function converts a string of time to a timestamp.

Syntax

TO_TIMESTAMP ( string )
Note:
The string parameter represents the input time string.

Sample

TO_TIMESTAMP('2007T')                         -- `2007T`
TO_TIMESTAMP('2007-02-23T12:14:33.079-08:00') -- `2007-02-23T12:14:33.079-08:00`

UTCNOW

The UTCNOW function returns the current timestamp in UTC.

Syntax

UTCNOW()

Sample

UTCNOW() -- 2019-01-01T14:23:12.123Z

String Functions

COS Select supports the following string functions:

CHAR_LENGTH, CHARACTER_LENGTH

Both CHAR_LENGTH and CHARACTER_LENGTH can compute the number of characters in a string, and they have the same semantics.

Syntax

CHAR_LENGTH ( string )
Note:
The string parameter specifies the string for character counting

Sample

CHAR_LENGTH('null')      -- 4
CHAR_LENGTH('tencent')   -- 7

LOWER

The LOWER function converts all uppercase letters in the specified string to lowercase letters, with all non-uppercase letters left unchanged.

Syntax

LOWER ( string )
Note:
The string parameter specifies the string for which to convert uppercase letters to lowercase letters.

Sample

LOWER('TENcent') -- 'tencent'

SUBSTRING

The SUBSTRING function returns a substring of a string. You can specify an index from which the SUBSTRING function will extract the remainder of the original string based on the length of the specified substring and return the result.
Note:
If the input string contains only 1 character, and the index is set to be greater than 1, the SUBSTRING function will automatically switch it to 1.

Syntax

SUBSTRING( string FROM start [ FOR length ] )
Note:
The string parameter specifies the string from which to extract a substring.
The start parameter represents an index value of the string as the starting position for extraction.
The length parameter specifies the length of the substring. If the length of the substring is not specified, the remainder of the string will be extracted.

Sample

SUBSTRING("123456789", 0)      -- "123456789"
SUBSTRING("123456789", 1)      -- "123456789"
SUBSTRING("123456789", 2)      -- "23456789"
SUBSTRING("123456789", -4)     -- "123456789"
SUBSTRING("123456789", 0, 999) -- "123456789" 
SUBSTRING("123456789", 1, 5)   -- "12345"

TRIM

The TRIM function deletes all characters before the first character or after the last character of the specified string. " " is the default character to be deleted.

Syntax

TRIM ( [[LEADING | TRAILING | BOTH remove_chars] FROM] string )
Note:
The string parameter specifies the string to be manipulated.
The LEADING | TRAILING | BOTH parameter specifies the extra characters to be deleted, which can be before the string (LEADING), after the string (TRAILING), or both (BOTH).
The remove_chars parameter specifies the type of extra characters to be deleted. It can be a string containing more than one characters. The TRIM function will delete all extra characters of the corresponding type that are identified by the TRIM function before or after the string parameter.

Sample

TRIM('       foobar         ')               -- 'foobar'
TRIM('      \tfoobar\t         ')            -- '\tfoobar\t'
TRIM(LEADING FROM '       foobar         ')  -- 'foobar'
TRIM(TRAILING FROM '       foobar         ') -- 'foobar'
TRIM(BOTH FROM '       foobar         ')     -- 'foobar'
TRIM(BOTH '12' FROM '1112211foobar22211122') -- 'foobar'

UPPER

The UPPER function converts all lowercase letters to uppercase letters with non-lowercase letters left unchanged.

Syntax

UPPER ( string )
Note:
The string parameter specifies the string to be converted to uppercase letters.

Sample

UPPER('tenCENT') -- 'TENCENT'

Reserved Words

Last updated：2024-03-25 15:28:24

In order to ensure normal operation and subsequent expansion, the COS Select feature has the following reserved fields such as function name, data type, and operator. You may use these fields to facilitate your SQL queries.
No.
Field
No.
Field
No.
Field
No.
Field
No.
Field
1
absolute
51
create
101
goto
151
octet_length
201
struct
2
action
52
cross
102
grant
152
of
202
substring
3
add
53
current
103
group
153
on
203
sum
4
all
54
current_date
104
having
154
only
204
symbol
5
allocate
55
current_time
105
hour
155
open
205
system_user
6
alter
56
current_timestamp
106
identity
156
option
206
table
7
and
57
current_user
107
immediate
157
or
207
temporary
8
any
58
cursor
108
in
158
order
208
then
9
are
59
date
109
indicator
159
outer
209
time
10
as
60
day
110
initially
160
output
210
timestamp
11
asc
61
deallocate
111
inner
161
overlaps
211
timezone_hour
12
assertion
62
dec
112
input
162
pad
212
timezone_minute
13
at
63
decimal
113
insensitive
163
partial
213
to
14
authorization
64
declare
114
insert
164
pivot
214
trailing
15
avg
65
default
115
int
165
position
215
transaction
16
bag
66
deferrable
116
integer
166
precision
216
translate
17
begin
67
deferred
117
intersect
167
prepare
217
translation
18
between
68
delete
118
interval
168
preserve
218
trim
19
bit
69
desc
119
into
169
primary
219
true
20
bit_length
70
describe
120
is
170
prior
220
tuple
21
blob
71
descriptor
121
isolation
171
privileges
221
union
22
bool
72
diagnostics
122
join
172
procedure
222
unique
23
boolean
73
disconnect
123
key
173
public
223
unknown
24
both
74
distinct
124
language
174
read
224
unpivot
25
by
75
domain
125
last
175
real
225
update
26
cascade
76
double
126
leading
176
references
226
upper
27
cascaded
77
drop
127
left
177
relative
227
usage
28
case
78
else
128
level
178
restrict
228
user
29
cast
79
end
129
like
179
revoke
229
using
30
catalog
80
end-exec
130
limit
180
right
230
value
31
char
81
escape
131
list
181
rollback
231
values
32
char_length
82
except
132
local
182
rows
232
varchar
33
character
83
exception
133
lower
183
schema
233
varying
34
character_length
84
exec
134
match
184
scroll
234
view
35
check
85
execute
135
max
185
second
235
when
36
clob
86
exists
136
min
186
section
236
whenever
37
close
87
external
137
minute
187
select
237
where
38
coalesce
88
extract
138
missing
188
session
238
with
39
collate
89
false
139
module
189
session_user
239
work
40
collation
90
fetch
140
month
190
set
240
write
41
column
91
first
141
names
191
sexp
241
year
42
commit
92
float
142
national
192
size
242
zone
43
connect
93
for
143
natural
193
smallint
-
-
44
connection
94
foreign
144
nchar
194
some
-
-
45
constraint
95
found
145
next
195
space
-
-
46
constraints
96
from
146
no
196
sql
-
-
47
continue
97
full
147
not
197
sqlcode
-
-
48
convert
98
get
148
null
198
sqlerror
-
-
49
corresponding
99
global
149
nullif
199
sqlstate
-
-
50
count
100
go
150
numeric
200
string
-
-

Data Types

Last updated：2024-03-25 15:28:24

Overview

COS Select supports a wide variety of primitive data types.
The data type directly supported by a compiler is called a primitive data type.

Data Type Conversion

COS Select uses the CAST function to determine the data type of your input data. In general, if you do not specify a data type through the CAST function, COS Select will treat the input data as string type.
For more information on the CAST function, see the CAST section in the SQL function documentation.

Supported Data Types

COS Select supports the following primitive data types:
Name
Description
Example
bool
TRUE/FALSE
FALSE
int, integer
An 8-byte signed integer
Value range: -9,223,372,036,854,775,808 - 9,223,372,036,854,775,807
100000
string
A UTF-8-encoded string with a character length in the range of 1 - 2,147,483,647
'xyz'
float
An 8-byte floating point
CAST(0.456 AS FLOAT)
decimal, numeric
A decimal value with a maximum precision of 38 decimal places and in the range of $-2^{31}$ - $2^{31}-1$
123.456
timestamp
A timestamp represents a certain moment and can be expressed with any precision. A timestamp in text format follows the W3C specification but needs to end with "T" (unless recorded in days).
When a fractional second is used, it should be accurate to at least one decimal place (with no upper limit on the number of decimal places).
The offset of local time can be expressed by the offset expressed by hours and minutes when compared to UTC, or "Z" can be used to express the offset of local time from UTC. The time offset does not need to be displayed when only the date is recorded.
CAST('2007-04-05T14:30Z' AS TIMESTAMP)

Operators

Last updated：2024-03-25 15:28:24

COS Select supports the following operators:
Operator Type
Operator
Logical operator
AND, NOT, OR
Comparison operator
<, >, <= ,>= , = , <>, !=, BETWEEN, IN, such as IN ('a', 'b', 'c')
Pattern matching operator
LIKE
Mathematical operator
+, -, *, %

Priority of Operators

The following table shows the priority of operators in descending order:
Operator/Element
Association
Use Case
-
Right
Unary subtraction
*, /, %
Left
Multiplication, division, and modulo
+, -
Left
Addition and subtraction
IN
-
Setting membership
BETWEEN
-
In the range of ()
LIKE
-
String pattern matching
<>
-
Less than and greater than
=
Right
Equivalence and assignment
NOT
Right
Logical NOT
AND
Left
Logical AND
OR
Left
Logical OR

Data Extraction

Last updated：2025-10-16 16:39:37

Overview

COS Select allows you to filter out desired data at the storage level, significantly reducing the amount of data transferred by COS, thereby lowering your usage costs, and improving data acquisition efficiency. In the COS console, you can extract data from individual files stored in buckets using the standard SQL templates we provide or by specifying statements that comply with syntax rules.
Note:
COS Select currently supports objects in JSON or CSV format in public cloud regions (Chinese mainland) and Virginia, and objects in Parquet format only in Beijing region.
Please make sure the file to be extracted complies with COS Select specifications. For more information on COS Select specifications, please see SELECT Overview.
COS Select in the console supports extraction of up to 40 MB data from a maximum of 128 MB files. To process larger files or extract more data, please use the API or SDKs.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Locate the bucket where the object resides and click the bucket name.
4. Click File List on the left sidebar.
5. Find the target object and click More > Extract on the right.
Note:
Currently, only objects in the STANDARD, STANDARD_IA, and INTELLIGENT TIERING storage classes can be extracted.
6. Configure the fields File Type, Header Field, Separator, Compression Format, and Export Format.
7. Click Select SQL Template.
8. In the pop-up window, select the template statement you want to use for extraction and click OK.
9. Edit the statement in the text box based on your actual file and then click Run SQL.
Now you can see the first 100 results in the text box at the bottom. To obtain the complete set of data, click Export Extraction Results.

Log Management

Logging Overview

Last updated：2025-10-10 09:50:01

Feature Overview

The logging feature allows you to record the access information of a source bucket and save it to a destination bucket as logs.
In the destination bucket, the log path is:
destination bucket/path prefix{YYYY}/{MM}/{DD}/{time}_{random}_{index}
Logs are generated every 5 minutes (one record per line). Each record contains multiple fields, and fields are separated by a space. The file size for a log file is up to 256 MB. If the amount of logs generated within these 5 minutes exceeds 256MB, your logs will be split into multiple log files. The logging fields supported are as follows:
No.
Field
Description
Sample
1
eventVersion
Log version
1.0
2
bucketName
Bucket Name
examplebucket-1250000000
3
qcsRegion
Request region
ap-beijing
4
eventTime
Event time (request end time, which is a timestamp in UTC+0 time zone)
2018-12-01T11:02:33Z
5
eventSource
Access domain name
examplebucket-1250000000.cos.ap-guangzhou.myqcloud.com
6
eventName
Event name, i.e., API operation name. For details, see Operation List.
UploadPart
7
remoteIp
Source IP
192.168.0.1
8
userSecretKeyId
User access KeyId
************************************
9
reservedFiled
Reserved field.
Displayed as -
10
reqBytesSent
Request bytes
83886080
11
deltaDataSize
Change in storage made by the request (in bytes)
808
12
reqPath
Requested file path
/folder/text.txt
13
reqMethod
Request method
put
14
userAgent
User UA
cos-go-sdk-v5.2.9
15
resHttpCode
HTTP response status code
404
16
resErrorCode
Error Codes
NoSuchKey
17
resErrorMsg
Error Message
The specified key does not exist.
18
resBytesSent
Bytes returned
197
19
resTotalTime
Total time used by the request (in milliseconds, i.e., the time between the last byte of the response and the first byte of the request)
4295
20
logSourceType
Source type of the log
USER (user access requests), CDN (CDN origin-pull requests)
21
storageClass
Storage Type
STANDARD，STANDARD_IA，ARCHIVE
22
accountId
Bucket owner ID
100000000001
23
resTurnAroundTime
Time used by the request server (in milliseconds, i.e., the time between the first byte of the response and the last byte of the request)
4295
24
requester
Visitor account.
If the visitor is the root account, the account format is root account UIN:root account UIN;
If the visitor is a sub-account, the account format is root account UIN:sub-account UIN;
If the visitor is a service account, the account format is granted root account UIN:role ID;
If the requester is anonymous, it is displayed as -.
100000000001:100000000001
25
requestId
Request ID
NWQ1ZjY4MTBfMjZiMjU4NjRfOWI1N180NDBiYTY=
26
objectSize
Object size, in bytes
808, if you use multipart upload, the objectSize field will only be displayed upon completion of the upload. During each part's upload, this field will display -.
27
versionId
Object version ID
Random string
28
targetStorageClass
Destination storage class, recorded for replication requests
STANDARD，STANDARD_IA，ARCHIVE
29
referer
HTTP referer of the request
*.example.com or 111.111.111.1
30
requestUri
Request URI
"GET /fdgfdgsf%20/%E6%B5%AE%E7%82%B9%E6%95%B0 HTTP/1.1"
"GET /fdgfdgsf%20/%E6%B5%AE%E7%82%B9%E6%95%B0 HTTP/1.1"
Notes:
If the signature is in the query_string form, it will be shown in the requestUri. Otherwise, the requestUri will not contain the signature information.
31
vpcId
VPC request ID
"0" (non-VPC) / "12345" (VPC, any non-zero string)
Note:
Currently, the log management function of COS supports the following regions: Beijing, Nanjing, Shanghai, Guangzhou, Chengdu, Chongqing, Shenzhen Finance, Shanghai Finance, Hong Kong (China), Singapore, Seoul, Bangkok, Tokyo, Silicon Valley, Virginia, Frankfurt.
The destination bucket and source bucket must reside in the same region.
The destination bucket that stores logs can be the source bucket itself (not recommended).
Currently, logs will be generated only when the bucket is accessed through XML APIs and XML API-based SDKs or tools, not via JSON APIs or JSON API-based SDKs or tools.
Depending on customer needs and business development, COS may add new fields to the access logs. Be sure to prepare for this when parsing the logs.
COS does not guarantee 100% accuracy of log data, for reference only, not as measurement or billing basis.

Enabling Logging

Using the Console

1. Log in to the COS console.
2. In the left navigation bar, click Bucket List to enter the bucket list management page.
3. Find the source bucket for which you want to enable log storage, click the Bucket Name, and enter the bucket management page.
4. On the left sidebar, select Logging > Logging, click Edit, and toggle on the feature.
5. Configure the configuration items of Logging as follows.
Destination Bucket: The destination bucket that stores the source bucket's logs must reside in the same region as the source bucket. We recommend you not set the source bucket itself as the destination bucket.
Path Prefix: A custom path prefix for the logs. If this field is left empty, logs will be delivered to the root directory of the destination bucket.
Service Authorization: When enabling the log storage feature for the first time, you need to authorize the CLS product service to deliver access logs to your bucket. Follow the prompts to complete the authorization.
6. After confirming that everything is correct, click Save.
Note:
After enabling the log storage feature, it is expected to take 1 hour to complete the first batch of log delivery. Subsequently, logs will be generated every 5 minutes, and the specific time of log delivery depends on the data volume.
7. Go to the destination bucket configured to view the generated logs.

8. After downloading the logs, you can view them in detail. For more information, see Logging Overview.
Note:
Currently, only the bucket owner possesses the permission to configure log storage. Other users will not see the Log Storage configuration option when accessing the console.

Notes

To enable Logging, you need to create a log role in the CAM console and grant the role permissions to read/write logs of the source bucket.
If Logging is disabled yet the role is not deleted, the role's read/write permissions on the source bucket's logs still exist.

Using the API

To enable logging for a bucket using APIs, follow the steps below:

1. Create a log role

Create a log role. For more information, see CreateRole.
Here, roleName must be CLS_QcsRole.
policyDocument must be:
{
    "version": "2.0",
    "statement": [{
        "action": "name/sts:AssumeRole",
        "effect": "allow",
        "principal": {
            "service": "cls.cloud.tencent.com"
        }
    }]
}

2. Grant the log role permissions

Grant the log role permissions. For more information about the API, see AttachRolePolicy.
Here, policyName is set to QcloudCOSAccessForCLSRole.roleName is the CLS_QcsRole in step 1, or the roleID returned when roleName was created.

3. Enable Log Management

Call the API to enable logging. For more information about the API, see PUT Bucket logging. Note that the destination bucket to store the logs should be in the same region as the source bucket.

Restrictions on Log Management

Last updated：2024-03-25 15:28:24

The log management feature is used to record the detailed access information of the specified source bucket and store the information in the specified bucket in the form of log files to facilitate bucket management.
Currently, the use limits of the log management feature are as follows:
Delivery frequency limit: A log is generated every 5 minutes.
Limit on the size of log file to be delivered: The log file to be delivered each time can be up to 256 MB. After this limit is exceeded, a new file will be delivered.
Format of the log to be delivered: Each record is saved as one line and can contain multiple fields separated by space.
Restrictions on the fields to be delivered: For more information on the fields to be delivered, see Log Management Overview.
Invalid fields: If a field in a log contains the - character, the field is an invalid record or the default record.

Content Recorded in a Log

The requests initiated by you to upload, download, or delete an object, to create or delete a bucket, or to modify the bucket configuration.
The requests generated by content delivery through CDN and data pull from the COS origin server.

Content Not Recorded in a Log

Offline origin-pull requests: After origin-pull is configured, if there are no objects in COS, the data will be downloaded from the origin server specified by you. This download operation is an offline origin-pull request and will not be recorded in the log.
Redirects within a static website: If you configure redirection in the static website feature, you may be redirected to another page when accessing index.html. Such redirects will not be recorded in the log.
Lifecycle transitions and deletions: If you configure the lifecycle to transition or delete objects after expiration, the operations of transitioning or deleting objects will be performed by the COS backend and will not be recorded in the log.
Operations of listing objects and uploading inventory reports: The inventory feature automatically lists all or specified objects in your buckets and delivers the generated inventory report to your buckets on a periodical basis. The operations of listing objects and delivering inventory reports will not be recorded in the log.
Operations of cross-region object replication: Cross-region replication involves getting an object from the source bucket and uploading it to the destination bucket. These operations are performed by the COS backend and will not be recorded in the log.
Object downloads in the COS Select feature: The COS Select feature allows you to extract an object which needs to be obtained from a storage device first. Object downloads are performed by the COS backend and will not be recorded in the log.

Using COS to Store Tencent Cloud Product Logs

Last updated：2024-03-25 15:28:24

Overview

The process of using Tencent Cloud products generates a massive number of logs, which can be used to help you analyze and make decisions for your business. For your convenience, you can store these logs persistently in COS, and then easily access them for analysis by using various APIs, SDKs or tools.
COS log storage offers the following benefits:
Persistent storage: COS provides stable, persistent storage that allows you to store logs in COS at a very low cost. You can access your logs anytime, anywhere for business analysis or decision-making.
COS Select: This feature allows for the simple extraction of logs stored on COS. You can combine log fields to make it easier to extract the information you need and reduce data download traffic.
Data analysis: You can analyze one or more COS logs using Sparkling, which then can be used to facilitate and provide a basis for important business decisions.

Shipping Logs

You can ship your Tencent Cloud product logs to COS in one of the following two ways:
Use the Tencent Cloud log shipping feature in which logs are directly shipped to COS. Products such as COS and CloudAudit (CA) support directly shipping logs to COS.
Use the CLS log shipping feature in which logs are first shipped to CLS and then transferred to COS for persistent storage.
The table below shows the particular log shipping method or methods currently supported by various Tencent Cloud products:
Product
Supports Direct Shipping to COS
Supports Shipping to CLS
CA
Yes
No
CLB
No
Yes
CKafka
Yes
No
API Gateway
No
Yes
SCF
No
Yes
TKE
No
Yes
CSS
No
Yes
TCB
No
Yes (however, TCB does not support shipping logs to COS through CLS)
COS
Yes
Yes for accounts on the allowlist. You can contact us to add your account to the allowlist.

Directly shipping logs to COS

The Tencent Cloud products outlined below allow you to ship logs directly to COS by setting log shipping rules as instructed in the product-specific documentation.
Product
Log Shipping Document
Log Shipping Interval
Log Shipping Path
CA
10-15 min
cloudaudit/customprefix/timestamp
CKafka
5-60 min
You can specify the interval
instance id/topic id/timestamp
COS
5 min
You can specify the path prefix. We recommend specifying an identifiable path, e.g. cos_bucketname_access_log/timestamp.
Note:
CKafka supports shipping the message data that it generates. However, to get logs on certain actions such as the creation of CKafka instances, you need to ship the CA logs.

Using CLS to ship logs to COS

The Tencent Cloud products outlined below allow you to ship logs directly to CLS for search and analysis. With CLS, you can further ship the logs to COS for persistent storage, thus reducing your costs and allowing for convenient offline analysis.
Product
Documentation
API Gateway
TKE
CSS
Click here to view
CLS can ship the following three types of logs to COS:
CSV-formatted logs: Logs formatted with comma-separated values. For more information, see Shipping CSV-Formatted Logs.
JSON-formatted logs: Logs in JSON format. For more information, see Shipping JSON-Formatted Logs.
Original logs: Logs in their original format, including those with full text in a single line, full text in multi-lines, and comma-separated values (for certain products). For more information, see Shipping Original Logs.
To ship logs from CLS to COS, you need to do the following:
1. Choose a Tencent Cloud product that meets your business needs and configure a logset and log topic as instructed in the log shipping documentation provided above to connect your business data to CLS.
2. With your particular business needs in mind, select the most suitable log type to ship to COS. We recommend that you use the product name as the path prefix to differentiate product logs, for example, you could use the following name for TKE logs: TKE_tkeid_log/timestamp.
3. Once the shipping rule is configured, you can configure COS event notifications for file uploads through SCF. You can perform further operations on logs shipped to COS based on these event notifications. For more information, see Event Notifications.

Log Analysis

Downloading logs locally for offline analysis

You can download logs to the local file system through the Tencent Cloud console, or by using various SDKs, APIs, and tools. To do so, follow the instructions in one of the following documents and replace each file path in the code with your log storage path:
Download Method
Document
Console
COSBrowser
COSCMD
Android SDK
C SDK
C++ SDK
.NET SDK
Go SDK
iOS SDK
Java SDK
JavaScript SDK
Node.js SDK
PHP SDK
Python SDK
WeChat Mini Program SDK
API

Analyzing logs using COS Select

You have the option to extract and analyze COS logs stored in CSV or JSON format directly by using the COS Select feature. It enables you to query desired log fields, which largely reduces the number of logs transferred by COS, thus lowering usage costs and improving data access speed. For more information on COS Select, see Select Overview.
Currently, you can access COS Select using the console or APIs.
Method
Instructions
Console
API

Setting Log Retrieval

Last updated：2024-04-07 17:21:13

Overview

Log retrieval is also known as the real-time log function. Cloud Object Storage (COS) provides this function to record various bucket-related request logs and retrieve and analyze log data in real time. You can quickly enable this function for buckets in the COS console to retrace exceptional events and locate faults.
Note
To enable the real-time log function for a COS bucket, enable Cloud Log Service.
When using the real-time log query function, you are charged by Cloud Log Service (CLS). For details about billing standards, see Billing Overview.
Currently, the log retrieval function of COS is available in the regions, including Beijing, Guangzhou, Shanghai, Chengdu, Nanjing, Chongqing, Hong Kong (China), Silicon Valley, Singapore, Mumbai, Frankfurt, Toronto, Shenzhen Finance, and Shanghai Finance. This function will be available in more regions. Please stay tuned for product updates.
COS does not ensure the accuracy of log data. Log data is for reference only and cannot be used as a basis for measurement and billing.

Directions

Activating Log Retrieval

1. Log in to the COS console.
2. On the left sidebar, click Bucket List to go to the bucket list page.
3. Locate the bucket for which you want to activate log retrieval, and click the bucket name to go to the management page.
4. In the left sidebar, choose Logging > Log Retrieval.
5. On the Log Retrieval page, if the real-time log function is not enabled for the current bucket, click Activate Now.
Note:
After the function is enabled, access logs of this bucket will be shipped to the log topic named cos-log-store in the same region of CLS. Authorization is required for the initial activation.


Activating Field-Based Statistical Analysis

Log retrieval enables you to perform quick statistical analysis based on fields without the need to enter query statements. For default supported fields, see Log Fields. For more information, see Quick Analysis.
Note:
If you want to perform statistical analysis on logs based on a specific field, enable statistics for that field in the following two ways.

Method 1: Enable Statistics on the Raw Logs Tab Page

1. Go to the Raw logs tab page.
2. Locate the target field and click enable now.


Method 2: Enable Statistics on the Index Configuration Page

1. Click Index Configuration to go to the index configuration page.

2. On the index configuration page, click Edit in the upper right corner, locate the target field, turn on the switch in Enable Statistics column, confirm the information, and click OK.

3. On the page that appears, view statistical analysis of the field. Click a number to automatically generate the corresponding retrieval and analysis statements and chart.


Entering Retrieval and Analysis Statements

Note:
Index Configuration is a necessary condition for retrieval and analysis with CLS. You can perform retrieval and analysis on log data only when the index configuration function is enabled. Index configuration is complex, so it's recommended that you use the suggested configurations in Modifying index configuration to fulfill most use cases. For more information, see Configuring Indexes.
For the prerequisites and user guide of retrieval and analysis statements, see Syntax and Rules. For analysis methods of bucket access logs, see COS Access Log Analysis.
A retrieval and analysis statement consists of Criteria and SQL statements. Enter the retrieval and analysis statement to retrieve and statistically analyze access logs.
Criteria: Specify the conditions that the logs need to match, so that logs that meet the conditions are returned. For example, use status:404 to retrieve application request logs with a response status code of 404. If the criteria are empty or *, all logs are displayed. For more information about syntax rules and examples, see Syntax and Rules.
SQL statement: Statistically analyze logs that meet the criteria, so that the analysis results are returned. For example, use status:404 | select count(*) as logCounts to count the number of logs with a response status code of 404. For more information about syntax usage and examples, see SQL Statement Syntax Rules.

1. On the Log Retrieval page, select Interaction mode(CQL) or Statement mode(CQL) for retrieval and analysis statements.
**Interaction mode(CQL)**: In this mode, you can click specified retrieval criteria and statistical analysis rules to automatically generate retrieval and analysis statements. This mode is highly user-friendly.

**Statement mode(CQL)**: In this mode, you can directly enter retrieval and analysis statements following the syntax rules. This mode is highly flexible.

2. Click the drop-down list to switch to Syntax rule: CQL. The options are CQL (recommended) and Lucene. For more information about syntax rules and examples, see Search Condition Syntax Rules.

3. Enter the retrieval and analysis statement, select the time range, and click the Search button to retrieve the access logs reported by the bucket to CLS.

Viewing Log Retrieval and Statistical Analysis Results

After a successful retrieval, you can view the retrieval and statistical analysis results of access logs on the Log retrieval page. The results are respectively displayed on the Raw logs and Chart tab pages, which can be quickly switched. Details of the two tabs are as follows:
Raw logs: When a retrieval and analysis statement only contains criteria, you can view the logs matching the criteria on the Raw logs tab page. The logs are sorted by time in descending order.
Chart: When a retrieval and analysis statement contains SQL statements, you can view the analysis results on the Chart tab page and the logs matching the criteria on the Raw logs tab page. In this case, you can compare and analyze the statistical results and raw logs.

Raw Logs


1.1 You can view the statistics information of logs within a specified time range under the current criteria in a bar chart.

1.2 By default, log data is displayed in original form. The display form can be customized.

1.3 Only the fields with Show next to them are displayed in the Log Data column. You can view or hide more fields associated with logs in the following three ways:
List Operation: In the left panel on the Raw logs tab page, locate the target field and click Show or Hide to complete the configuration.

Log Details: Expand details of a log, and click the view icon to quickly hide or show a certain field. You can also view data of each log in JSON format and quickly copy the data. In addition, you can click a field value for custom redirection.

Layout: Click "Layout: Default configuration" on the right, and select Manage configuration. You can show fields in batches, and click Application to complete the configuration.



Chart

The Chart tab page shows analysis results in tables by default. You can customize the chart type and related information in Chart Configuration on the right. For further visualization and alarm configuration of bucket access logs, go to the CLS console.


Downloading Logs and Exporting Analysis Results

Downloads Logs

1. Click Download on the right and select Download Logs.

2. In the popup window, configure the information on the figure below, verify the configuration, and click Export.

3. In the Export Logs window where the new export task is displayed, view the current task progress, or delete or download the task. When the task is in the Waiting state, you can exit the window. Then you can click Download on the right and select Export Logs to enter the window again.


Exporting Analysis Data

Click Export Analysis Data on the right to download the analysis results to your local path.


Disabling Log Retrieval

If you do not want to use the log retrieval function for a bucket, disable the function.
Note:
Before disabling this function, go to the CLS console to check for empty log topics. If you do not delete them in time, extra charging will be incurred.

Log Field Description

Serial Number
Field Name
Description
Example
1
eventVersion
Record version
1.0
2
bucketName
Bucket name
examplebucket-1250000000
3
qcsRegion
Region for the request
ap-beijing
4
eventTime
Event time (end time of the request, which is a timestamp in UTC+0 time zone)
2018-12-01T11:02:33Z
5
eventSource
Domain name to be accessed
examplebucket-1250000000.cos.ap-guangzhou.myqcloud.com
6
eventName
Event name
UploadPart
7
remoteIp
Source IP address
192.168.0.1
8
userSecretKeyId
Key ID for access
AKIDNYVCdoJQyGJ5brTf
9
reqBytesSent
Number of bytes in the request
83886080
10
deltaDataSize
Change in storage made by the request (in bytes)
808
11
reqPath
File path for the request
/folder/text.txt
12
reqMethod
Request method
put
13
userAgent
User agent (UA)
cos-go-sdk-v5.2.9
14
resHttpCode
HTTP return code
404
15
resErrorCode
Error code
NoSuchKey
16
resErrorMsg
Error message
The specified key does not exist.
17
resBytesSent
Number of bytes in the response
197
18
resTotalTime
Total time consumed for the request, which is the time between the first byte of the request and the last byte of the response, in milliseconds
4295
19
logSourceType
Type of the log source
USER (user access request) and CDN (CDN origin-pull request)
20
storageClass
Storage class
STANDARD, STANDARD_IA, and ARCHIVE
21
accountId
Bucket owner ID
100000000001
22
requester
Requester account
The value is in the format of root account ID:sub-account ID. In case of an anonymous access, the value is shown as -.
23
requestId
Request ID
NWQ1ZjY4MTBfMjZiMjU4NjRfOWI1N180NDBiYTY=
24
objectSize
Object size, in bytes
808. If you use multipart upload, the objectSize field will be displayed only when the upload is completed. During the upload of each part, the field will be displayed as -.
25
versionId
Object version ID
Random string
26
targetStorageClass
Target storage type, which is recorded for any replication request
STANDARD, STANDARD_IA, and ARCHIVE
27
referer
HTTP referer of the request
*.example.com or 111.111.111.1
28
requestUri
Request URI
"GET /fdgfdgsf%20/%E6%B5%AE%E7%82%B9%E6%95%B0 HTTP/1.1"
29
resTurnAroundTime
Time consumed by the request server, which is the time between the last byte of the request and the first byte of the response, in milliseconds
4295
30
vpcId
VPC Request ID
"0": non-VPC; "12345": VPC, which is a non-zero string

Common Examples

A few commonly used examples are provided below. More cases will be added later. Please stay tuned for product updates. For more information, see Examples.

Example 1: Top 50 IP Addresses by Traffic Volume

1. Select Statement mode(CQL) and enter the following SQL statement:
  * | select count(*) as PV, remoteIp as IP  group by IP order by PV desc limit 50
count(*) as PV: Counts all log entries, which is PV.
group by IP: Groups by IP, counting the PV for each IP.
order by PV desc: Sorts by PV in descending order, prioritizing IP addresses with high PVs.
limit 50: Only returns the top 50 query results, which are the IP addresses with the top 50 high PVs.

2. Select the time range and click Search. On the statistical analysis page, view the analysis results. You can switch chart types on the right side.

3. If the query statement fails to run, check the error location specified in the error message. For more information about error reasons, see Search Analysis Error.

Example 2: Querying Files of Access Within Recent 90 Days

1. Select Statement mode(CQL) and enter the following SQL statement:
* | select reqPath group by reqPath limit 1000000

2. Set the time range to the recent 90 days and click Search. On the statistical analysis page, view the analysis results. You can switch chart types on the right side.


Example 3: Counting Requests, Traffic, and Storage for Specified Prefixes

For example, all the following files within the bucket generate log requests. You may desire to count the requests, traffic volume, and storage volume for objects prefixed with 1(that is, the objects with file paths "/111", "/112", and "/101").

Note:
Object traffic: Total number of object bytes returned for log requests.
Object storage volume: Sum of object sizes (objectSize) in log requests, excluding "-". For more information, see Log Field Description.
1. Select Interaction mode(CQL), click Add criteria, specify reqPath to contain /1, and set the time range to filter object request logs prefixed with 1, where is a wildcard.

2. Enter the following statement in the input box:
SELECT count(*) as "Number of requests", sum(resBytesSent) as "Traffic volume", SUM(CAST(objectSize AS INT)) AS "Storage volume" WHERE objectSize != '-'

3. Alternatively, select Statement mode(CQL) in Step 1 and directly enter the following SQL statement:
reqPath:"/1*" | SELECT count(*) as "Number of requests", sum(resBytesSent) as "Traffic volume", SUM(CAST(objectSize AS INT)) AS "Storage volume" WHERE objectSize != '-'

4. Select the time range and click Search. On the statistical analysis page, view the analysis results. You can switch chart types on the right side.


Example 4: Locating Reasons for File Access Failure

1. Select Interaction mode(CQL) and click Add criteria. Specify the target file name, select the Time range and click Search.

2. Click the field name on the left, enable quick analysis, and conduct statistical analysis based on the resHttpCode field. View log information of error codes, such as 403 and 204.

3. Click an error code and choose Add to Search to filter the logs. In the log details, view information about other fields, focusing on the fields such as reqMethod. The figure below shows that the delete operation is performed on the file, resulting in object access failure.






Setting Log Analysis

Last updated：2025-06-25 16:59:40

Overview

If COS log storage is enabled, you can use the log analysis feature to further analyze the generated log files. This feature consolidates log files within a specified time range for statistical analysis and extracts key metrics for your reference.

Prerequisites

Log storage feature is enabled. If not enabled, please follow the Set Log Storage guide to enable it.
Note:
The applicable regions for the COS Log Analysis Function depend on the log storage feature. Please first enable the log storage feature before using the log analysis function.

Directions

1. Log in to the COS console.
2. Choose Bucket List on the left sidebar.
3. Find the bucket that needs log analysis. Click the bucket name to go to the bucket management page.
4. On the left sidebar, select Log Management > Logging.
Note:
To use the log analysis feature, first enable log storage as instructed in Setting Logging.
5. With the log storage service enabled, the activation page for log analysis will appear. Click Use Now, and the system will detect whether you have added a COS log analysis function. A pop-up will guide you to create a COS log analysis function. For configuration instructions, please refer to Add COS Log Analysis Function.
6. Select a corresponding function and time range. Click New Analysis Task and configure the following information in the pop-up window:
Source Bucket: The default is the current bucket.
Time Range: Period of logs you want to analyze, up to 30 days. Logs are retrieved according to end time. Logs in the specified time range cannot exceed 200 GB.
Cloud Function: Select a COS log analysis function added in the region where this bucket resides.
Product Destination Directory: After analysis, the output result will be zipped and saved to the directory you set. The output file contains result file and inventory file. Result file is based on the scenario you select. Inventory file refers to the log file list retrieved for this analysis.
7. Click Next, enter the detailed configuration page. Configuration instructions are as follows.
Task Type: Supports statistical analysis and retrieval operation logs.
Statistical analysis: Perform statistical analysis in scenarios that support object download.
Search operation logs: Support search related logs by feature operations
Statistical Analysis
Scene categories: Currently only support object download scenarios.
Select scene: Currently supported log analysis scenarios. For example, analyze the top N IPs and top N files with the highest download traffic, the top N IPs, top N files, top N HTTP referers, and top N UserAgents with the most download counts.
Value of N: Corresponding N value in the scenario. Enter a positive integer.
Job Description: Customize the description for this analysis.
Search operation logs
Scene categories: Support bucket and object-related scenarios.
Select Scene
Bucket-related: Including static website-related, origin-pull settings-related, lifecycle-related, tag management-related, inventory setting-related, cross-domain configuration-related, anti-hotlinking setting-related, server encryption-related, bucket ACL-related, Policy permission setting-related, global acceleration domain name setting-related, version control-related, cross-bucket replication-related, log storage-related.
Operation Filtering: Supports filtering read and write operations.
Object-related: Supports file upload (single file upload, multi-part upload, file replication), file download, single file deletion, and batch file deletion scenarios.
Filter Condition: Supports specifying object keys (not supported for batch file deletion), UserAgent, and source IP filter conditions.
Filter Mode: Supports fuzzy matching and exact matching.
Job Description: Customize the description for this analysis.
8. After confirming that the configuration is correct, click OK, and you will see that the analysis task has been added successfully. If the configuration is incorrect, follow the prompt to complete the operation.
Click View Result to view the result of this analysis task and where the result files are saved.
Note:
You can query log analysis tasks created in the last 3 days.
You can view the result only after a log analysis task is finished. A task takes a few to tens of minutes depending on the log size.
Click Running Log to go to the SCF console and view logs of COS log analysis.

Batch Operation

Overview

Last updated：2024-03-25 15:33:39

With the batch operation feature in COS, you can specify an operation on a specified list of objects in a bucket. To do so, you have two options: either create an inventory of the objects via the inventory feature or list the desired objects in a CSV file as formatted in the inventory. Those objects will then be batch operated as specified in COS.
For more information, see Inventory Overview.
At present, the batch operation feature only supports the following operations:
Batch copying objects
Batch restoring archived objects
You can use the batch operation feature in the COS console. For more information, see Batch Operation.

How It Works

To perform a batch operation, you need to create a batch operation job first, which contains all the information needed to perform the specified operation on the object list. You can use an inventory as an object list.
After you provide an inventory file and start the created batch operation job, the batch operation feature will execute the specified operation on the objects in the inventory sequentially. During job execution, you can monitor the execution status in the COS console or choose to output a job report after the job is completed. The job report provides details about each operation in the job.
Note:
The batch operation feature is only applicable to objects in the current bucket. If you want to batch operate on the objects in another bucket, please enable the batch operation feature for that bucket.

Object Inventory

An object inventory is a list of all the objects to be operated on. To create a batch operation job, you need to provide an object inventory first to tell COS what objects it should operate on. You need to put the object inventory file in the bucket and provide information such as the filename, ETag, and VersionID (if applicable). You can create an object inventory in the following two ways:
COS inventory feature: This feature outputs an object inventory in CSV format. For more information, see Inventory Overview. If your object inventory contains version ID information, COS will batch operate on the objects with the corresponding version ID.
CSV file configuration: Every row in the CSV file must contain the bucket name and the name and version ID (if versioning is enabled for the bucket) of an object for batch operation. If versioning has never been enabled, you can skip the version ID information. The CSV file can be configured as follows:
examplebucket-appid, exampleobject, PZ9ibn9D5lP6p298B7S9_ceqx1n5EJ0p
examplebucket-appid, exampleobject, jbo9_jhdPEyB4RrmOxWS0kU0EoNrU_oI
Note:
If versioning is or was once enabled for your bucket, and you want to perform a batch operation on the specified version of the object, you must provide object version ID information in the object inventory.
If versioning is or was once enabled for your bucket, but you don't specify a version ID in the inventory, COS will operate on the latest version of the object by default.
If you uploaded an object with the same name as an object to be operated on before creating a job, COS will operate on the object on the latest version by default rather than the version when the object inventory was created. To avoid this issue, you can enable versioning and specify the version ID in the object inventory.
An object inventory can contain all the objects in a bucket. However, note that it takes longer to operate on a large number of objects.

Batch Operation Job

This section describes how to create a batch operation job and how the system will respond after creation.
When creating a batch operation job, you need to provide the following information:
Type
Description
Operation
You need to specify which operation is to be performed on objects. Corresponding parameters can be configured for each operation, and COS will perform the operation on the objects in the inventory sequentially as configured.
Object inventory
An object inventory is a list of all the objects to be operated on. You can create an object inventory through the inventory feature. For more information, see Inventory Overview. You can also list desired objects in a CSV file as formatted in the inventory.
Priority
You can set priority to identify the precedence of the current batch operation job over other jobs. Job priority does not directly determine the order in which jobs will be completed. If you want to control the order of multiple jobs, you need to check the job execution status on your own and start the next job after the current one is completed.
Rule permission
After creating a batch operation job, you need to make sure that your account has the corresponding IAM permission to perform the operation. For example, if you have created a batch operation job to execute PUT Object-copy, you should make sure that your account has the Get Object permission in the source bucket and the PUT Object permission in the destination bucket. In addition, for all batch operation jobs, you should have permission to read the object inventory and write into the job report. For more information on permission configuration, see Overview and Bucket Policy.
Job report
If you want to output a job report after a batch operation job is completed, you need to enter the corresponding parameters when creating the job, so that the system can correctly output the job report to the specified destination bucket. The required information includes the bucket to store the job report, job report format, and whether all job information should be included. The file prefix of the job report is optional.
Job description (optional)
You can enter 256-byte job description for your created batch operation job to keep track of it. The job description will be displayed in the COS console and can be used to sort or filter jobs. You can enter same job description for similar jobs (e.g., syncing and copying log data weekly) to manage them in a centralized manner.

Performing Batch Operation

Batch Operation

Last updated：2025-05-23 18:37:23

Overview

The batch processing feature of Cloud Object Storage (COS) enables large-scale batch operations on objects in a bucket. It also supports using features such as audio and video transcoding and speech recognition from Cloud Infinite (CI) to perform batch processing on specified files within the bucket. Currently, you can perform batch processing on the following operations:
Copying objects
Restoring archived objects
audio and video transcoding, video matting, text to speech, and other data processing
You can generate an inventory file for the objects on which to perform a batch operation by using either COS inventory (you need to first enable the inventory feature as instructed in Setting Inventory), or the CSV format you specify. COS will then perform this batch operation based on the inventory file. For more information on batch operations, see Batch Operation Overview.

Directions

1. Log in to the COS console.
2. On the left sidebar, click Batch Operation to enter the batch operation management page.
3. Click Create Job to create a batch operation job.
Task name: Custom name for batch tasks.
Batch operation type: Select COS Batch to perform basic operations such as copy and archive, or select Standalone tasks and workflows to use related services of Cloud Infinite.
Job type: When COS Batch is selected, options include batch data copy, batch restore archived objects, and batch delete.
Inventory report
(1) If there is no inventory report: Select none. The system will generate an instant inventory in the selected inventory bucket. You can set inventory information as needed, such as file scope, object version, filtering tag, and filtering time. For more configuration item descriptions, please see Setup Checklist.



(2) If there is an inventory report or CSV: select Yes, and then select the bucket where the inventory is located.



Source Bucket: Select the bucket to store the inventory file.
Note：
Currently, COS batch operation is only available in public cloud regions in Chinese mainland, Silicon Valley region,Sao Paulo and Singapore.
Inventory Format: Select a format for the objects to be inventoried from the following two options:
Inventory Format
Field
Configuration Description
COS inventory report
-
You can use an inventory report generated by COS as the inventory file or create your own inventory file in CSV format.
CSV
Bucket
Bucket name
Key
Name of the object in a bucket. The object name encoded in URL format must be decoded and then can be used properly.
VersionId
Object version ID. If versioning is enabled for a bucket, COS will assign a version ID to each object added to the bucket. If you don't want to inventory the latest version of an object, you can specify a particular version ID.
Inventory File Path: Specify the path of COS inventory report or CSV file in the format: `directory/manifest.json` or `directory/manifest.csv`, respectively. For example, if you have an inventory file stored in the `examplebucket-1250000000` root directory, the inventory path will be `manifest.json`.
Inventory ETag: After selecting the inventory file path, it is automatically generated by the system. ETag (entity tag) is the hash of an object. The ETag only reflects changes to the content of an object, not changes to its metadata. The ETag is not necessarily the MD5 summary of the object data; this depends on how the object is created and the encryption method.
4. Click Next, and configure according to the configuration item description.



If Batch Data Copy was selected in the previous step, the configuration items are described as follows:
Destination Bucket: Select the bucket to store the object copies.
Prefix Operation: You can choose to add, replace, or delete the prefix on the object copies.
Storage Class: Specify the storage class for object copies. Valid values: STANDARD; STANDARD_IA; ARCHIVE; DEEP ARCHIVE.
Server-Side Encryption: Specify whether to encrypt the object copies. Valid values: None; SSE-COS.
Access Permission: Set access permissions to the object copies. Valid values: Copy all permissions; Replace all permissions; Add new permissions.
Object Metadata: Configure metadata for the object copies. Valid values: Copy all metadata; Replace all metadata; Add new metadata.
Object Tag: Configure tags for the object copies. Valid values: Copy all tags; Replace all tags; Add new tags.
If Batch Restore Archive Type Objects was selected in the previous step, the configuration items are described as follows:
Restoration Mode: You can select either standard or bulk mode. For more information on restoration modes, see Restoring Archived Objects.
Validity: Specify the number of days after which the object copies will expire and be automatically deleted. Value range: 1-365.
If Batch Delete was selected in the previous step, proceed to Step 5.
5. Click Next and configure the following options.



Job Priority: A job of a higher priority will be performed first. The value must be a positive integer. A larger value indicates a higher priority.
Job Report: Select whether to generate a job report.
CAM Role: You can create a CAM role or select an existing role to grant operation permissions to COS.
Note:
For COS to perform batch operations, you need to use a CAM role to grant permissions. For more information on CAM roles, see Role Overview.
6. Click Next, check the batch operation job configuration you set, and select If you select this option, the job will be directly started after creation. Make sure that the above configuration information is correct. as needed. To modify the information, click Modify or Previous.


7. After confirming that everything is correct, click Create or Create and Start. Once completed, you can find the new job in the job list. If you want to cancel the job, click Cancel Job under Operation.




Batch Copying Objects

Last updated：2024-03-25 15:33:39

The batch copy feature is used to copy objects in the inventory, i.e., allowing you to batch copy the specified objects from the source bucket to the destination bucket in the same region or in a different region. It supports customizing parameters for the PUT Object-copy operation, and parameter configuration affects the metadata and storage class of the target objects. For more information, see PUT Object-copy.

Restrictions

All objects to be copied must be in the same bucket.
Only one destination bucket can be configured for a batch copy job.
You need to have permission to read objects from the source bucket and write objects into the destination bucket.
The total size of the objects cannot exceed 5 GB.
Verification through ETags and server-side encryption using custom keys are not supported.
If the destination bucket does not have versioning enabled and contains an object with the same name as an object to be copied, COS will overwrite the object.
If an object in the inventory has multiple versions, only one version can be copied. If the version ID is not specified, the latest version will be copied.

Batch Restoring Archived Objects

Last updated：2024-03-25 15:33:39

The batch restoration feature can be used to restore archived objects in the inventory. This operation supports configuring parameters for POST Object restore. The configuration information will affect the restoration time and expiration time of copies. For more information, see POST Object restore.
Restoration is supported for ARCHIVE and DEEP ARCHIVE storage classes. For more information on these two storage classes, see Storage Class Overview.
To create a batch restoration job, you need to specify the following two parameters:
Restoration mode: Standard retrieval or bulk retrieval. For more information, see Restoring Archived Objects.
Copy validity period: When an archived object is restored, a temporary copy will be created, which will expire and be deleted automatically after the specified time period. For more information on the copy validity period, see Restoring Archived Objects.

Notes

1. If the batch restoration job includes copies of objects that have already been restored, the validity period of these copies will be updated when the job starts, ensuring that all copies in the job have the same validity period.
2. The batch restoration job only initiates requests to restore objects. After all requests are sent, the Batch Operation page will display the job as completed. COS doesn't notify you when the objects have been restored, but you can configure event notifications to receive notifications. For more information, see Event Notifications.

Batch Delete Objects

Last updated：2025-12-10 17:14:59

Batch delete operation can be used to delete objects in the inventory. For instructions on the delete operation, refer to DELETE Object. The version control status of the bucket affects the execution result of batch delete. For details, refer to version control overview.

Note

1. Based on whether the inventory includes the object's versionid, it determines whether the delete request initiated during the batch delete task includes the versionid.
When the inventory only retains the current version, it will perform a delete operation without versionid, namely DELETE /example.txt
When the inventory lists historical versions, it will perform a delete operation with versionid, namely DELETE /example.txt?versionid=xxx
2. When the bucket is in one of three different states—versioning not enabled, versioning suspended, or versioning enabled—a delete operation without versionid will yield different results. However, a delete operation with versionid ensures the object is permanently deleted. For details, refer to Versioning Overview. The behavior of batch deleting objects under different conditions is as follows:
Operation
List 1 (Keep Current Version Only)
Inventory 2 (Listing Historical Versions)
delete operation
without versionid
with versionid
Versioning not enabled
permanent deletion
permanent deletion
Versioning enabled
Adding a delete marker does not result in permanent deletion.
Performing permanent deletion using the versionid
Suspending Versioning
Adding a delete marker, if there is a null version object in the bucket, that object will be permanently deleted, and non-null version objects will not be deleted.
Performing permanent deletion using the versionid

Managing Batch Operation Jobs

Last updated：2025-11-24 17:46:11

This document describes how to manage batch operation jobs in the COS console.

Filtering Jobs

You can view the list of the batch operation jobs created in the last 90 days using the List Jobs API. The list contains information of each job, such as job ID, description, priority, status, and execution progress. You can filter them by job status in the list to display the ones in the same status or filter them by job description or ID in the console.

Querying Job Status

If you need more information, you can get all the information of a job using the DescribeJob API. This API will return information such as the task configuration of the specified job, object inventory information, and job report.

Setting Job Priority

You can set priority for your batch operation jobs, and COS will execute them based on priority, i.e., executing jobs of a higher priority first. Priority is expressed as an integer; the greater the value, the higher the priority. You can modify the priority during job execution. If a higher-priority job needs to be added, you can run it first by pausing the lower-priority job.
Note:
Although higher-priority jobs generally take precedence over lower-priority ones, priority is not a criterion for sequential execution. If you want to run batch operation jobs sequentially, please monitor the execution status of each job on your own and start them manually.

Job Status

Job status changes as the execution proceeds, as shown below:



The specific meaning of each job status is as follows:
Job Status
Description
Next Status
New
When a job is created, its status will be "New".
It can change to "Preparing", meaning that COS starts to parse the inventory in the job.
Preparing
When COS starts to parse the configured inventory in the batch operation job, the job status will change to "Preparing".
It can change to "Ready" or "Suspended". The former means that COS has finished parsing the job configuration information and is going to perform the specified tasks on the objects in the inventory based on the configuration, while the latter means that COS has finished parsing the job configuration information but is still waiting for your confirmation before it proceeds to perform the specified tasks (this status appears if a batch operation job is configured in the console).
Suspended
A batch operation job pending your confirmation is in the "Suspended" status, which appears if the job is created in the console. After your confirmation, COS will start executing the job, and the status will change accordingly and cannot revert to "Suspended".
When you confirm to execute the job, the job status will change to "Ready".
Ready
When COS has finished parsing the object inventory and job configuration information in your batch operation job and is going to perform tasks, the job status will change to "Ready".
It can change to "Active", at which point COS starts executing the job. If a higher-priority job is running, COS will keep the job status as "Ready" until the job status of the higher-priority job changes to "Complete".
Active
When COS is performing tasks on the objects in the inventory based on the configuration information, the job status will be "Active". You can query job progress in the console or by calling the DescribeJob API.
It can change to "Complete", "Failing", "Pausing", or "Canceling" if your job succeeds, fails, or is paused or canceled.
Pausing
After COS pauses an ongoing job and before the status changes to "Paused", the job status in between will be "Pausing".
It can be followed by "Paused" when COS has successfully paused the ongoing job.
Paused
If a higher-priority job is created, the status of the ongoing job will change to "Paused".
If the higher-priority job is completed, fails, or is pending confirmation, the job in "Paused" status can automatically shift to "Active" status.
Complete
When the batch operation job has successfully completed the tasks on all the objects in the inventory or fails, the job status will change to "Complete".
If job report generation is configured, COS will deliver the report to your specified bucket when the status changes to "Complete".
"Complete" is a final status, i.e., the job will not change to any other status.
Canceling
After COS cancels an ongoing job and before the status changes to "Canceled", the job status in between will be "Canceling".
It can be followed by "Canceled" when COS has successfully canceled the ongoing job.
Canceled
When a batch operation job is successfully canceled, the job status will change to "Canceled". At this point, you cannot make any modification to the job status.
"Canceled" is a final status. i.e., the job will not change to any other status.
Failing
The "Failing" status comes before "Failed".
It can change to the "Failed" status.
Failed
After a job fails, the job status will change to "Failed". For more information, see Tracking Job Failure.
"Failed" is a final status, i.e., the job will not change to any other status.

Tracking Job Failure

If a problem arises during job execution, such as trouble in parsing the object inventory, the batch operation job will fail, and COS will return the corresponding error code and cause. You can call the DescribeJob API or view the job report to get the cause of failure and other information.
COS sets a task failure threshold for each batch operation job to avoid frequent task failures. If a job involves more than 1,000 tasks, COS will monitor the task failure rate, i.e., proportion of failed tasks to all tasks executed. If the rate exceeds the threshold of 50% at any moment, COS will terminate the job and return the failure message. You can check the cause why this happens (e.g., the object inventory contains large amounts of information on non-existing objects). Then, you can fix the problem accordingly and create a job again.
Note:
COS executes batch operation jobs in an async manner and does not necessarily perform tasks in the same order that the objects are listed in the inventory. Therefore, you cannot determine which object is being operated on according to the order in the object inventory and whether a task is successful or fails. However, you can get the information on successful or failed tasks from the job report.

Job Report

You can configure whether to output a job report when creating a job. If yes, COS will output a job report when the job succeeds, fails, or is canceled. You can view the information on all successful or failed tasks in the job report.
A job report contains information such as the configuration parameters and execution status of the specified task as well as name and version ID of the objects operated on, task status codes, and error description.

Global Acceleration

Overview

Last updated：2025-10-29 16:13:02

The global acceleration feature provided by Tencent Cloud Object Storage (COS) utilizes a load balancing system based on Tencent's global traffic scheduling to intelligently route and parse user requests and select the optimal network linkage for nearby access. Backed by globally deployed Tencent Cloud data centers, it allows users across the world to quickly access buckets and improves application access success rate, allowing for greater business stability and a smoother user experience. In addition, COS's global acceleration also speeds up uploads and downloads.
Note:
Global acceleration supports COS public cloud regions. However, it is not supported in finance cloud regions due to network isolation.
Using this feature incurs fees as request data transfers are accelerated via Direct Connect lines in the Tencent Cloud private network. For more pricing information, see Product Pricing.

Directions

You can enable global acceleration on the COS Console or through APIs.

Using the COS console

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to go to the bucket list page.
3. Locate the bucket that requires global acceleration configuration, click its bucket name to enter the bucket details page.
4. In the left navigation bar, select Domains and Transfer > Global Acceleration, then click Edit in the Global Acceleration section to enable the current status.
5. After confirming that everything is correct, click Save to enable the bucket's global acceleration.
Global acceleration domain name: When uploading objects via this domain, public network global acceleration effect is achievable. The domain name format is <BucketName-APPID>.cos.accelerate.myqcloud.com.
6. After enabling global acceleration, you only need to access the bucket via the global acceleration domain name to achieve quick data access.
Note:
Note: Enabling global acceleration will not affect the original default bucket domain name. You can still use it normally.

Using RESTful APIs

You can directly use the following APIs to enable global acceleration:
PUT Bucket Accelerate
GET Bucket Accelerate

Access Endpoint Domain Names

After enabling global acceleration, you can access your COS files through two types of endpoint domain names:
Default bucket endpoint domain name: Format: <BucketName-APPID>.cos.<Region>.myqcloud.com. For more information, see Regions and Access Endpoints.
Global acceleration endpoint domain name: Format: <BucketName-APPID>.cos.accelerate.myqcloud.com.
Take the bucket examplebucket-1250000000 in Guangzhou as an example. If you have enabled global acceleration for it, when you need to upload the file exampleObject.txt from Beijing to it, you can do so in the following two ways:
Use the global acceleration endpoint domain name for access: when uploading the object, you need to set the endpoint domain name to examplebucket-1250000000.cos.accelerate.myqcloud.com. When you upload the object through this endpoint domain name, COS will intelligently parse your request based on your network conditions and implement nearby access. For example, it will forward your request to the Beijing access layer and then transmit it to the Guangzhou storage layer via a private network Direct Connect line to accelerate data transfer.
Use the default bucket endpoint domain name for access: when uploading the object, you need to set the endpoint domain name to examplebucket-1250000000.cos.ap-guangzhou.myqcloud.com. When you upload the object through this endpoint domain name, your request will be directly forwarded to the Guangzhou access layer and then the Guangzhou storage layer. In this case, long public network linkage may lead to unstable transfer.
Note:
If you use global acceleration, fees will be incurred. Therefore, we recommend that you carefully evaluate whether to use this feature with your actual business needs in mind:
If your business has more writes (e.g., PUT Object, POST Object, and Multipart Upload) than reads and uploads data to Tencent Cloud data centers from a remote region, we recommend using a global acceleration endpoint domain name.
If your business has more reads (e.g., GET Object) than writes and mainly involves file download, we recommend that you perform a comprehensive evaluation of the CDN-based access acceleration solution and select the option with the best cost performance.
If your business mainly involves configuration operations or file extraction, we recommend using the default bucket endpoint domain name.
If your business needs to access buckets over a private network or through a Direct Connect line within the same region, we recommend using the default bucket endpoint domain name.

Configuring DNS Intelligent Resolution

f businesses need worldwide access, they can configure DNS intelligent resolution to determine if a user is in domestic or outside Chinese mainland for faster response to user access. Go to the appropriate DNS service provider to configure the CNAME resolution record. Main directions:
1. Add domain: Add a custom domain name that requires intelligent parsing, then go to its record management page.
2. Add record: Create 2 CNAME records (for example, set the host record to accelerate). Set the resolution record type of the first entry to domestic and the record value to the bucket's default domain name examplebucket-1250000000.cos.ap-guangzhou.myqcloud.com. Set the resolution record type of the second entry to default and the record value to the global acceleration domain name examplebucket-1250000000.cos.accelerate.myqcloud.com.
3. Log in to the COS console and configure a custom domain name for the bucket, selecting the global acceleration origin site as the origin server. A red exclamation mark may appear next to the custom domain name, indicating a domain resolution issue, which can be ignored.
4. After configuration as per the aforementioned method, access your custom domain name to achieve the effect:
If your local DNS egress IP originates from outside Chinese mainland, the domain names or IP addresses queried and retrieved will be the global acceleration domain.
If your local DNS egress IP comes from domestic, the domain names or IP addresses queried and retrieved will be the bucket's default domain.
5. If you use the COS SDK, you can initialize the client with a custom domain name to achieve the desired effect. Take JS SDK as an example:
import COS from 'cos-js-sdk-v5';

const cos = new COS({
    SecretId: 'your_tmpSecretId', // temporary secretId issued by the sts service
    SecretKey: 'your_tmpSecretKey', // temporary secretKey issued by the sts service
    SecurityToken: 'your_sessionToken', // temporary SessionToken issued by the sts service
    Domain: 'your.custom.domain' //input custom domain name
 });

Notes

We have outlined below some important factors to note when using a global acceleration endpoint domain name:
The global acceleration domain name will take effect 15 minutes after being enabled.
After the global acceleration domain name is enabled, the maximum bandwidth for a bucket will be allocated based on the business volume of the entire network.
After the global acceleration domain name is enabled, only requests using that domain name will be accelerated. However, the default bucket domain name can still be used.
When using a global acceleration domain name, fees will be incurred only for requests for which linkage is accelerated. For example, if you use a global acceleration domain name to upload data from Beijing to a bucket in Beijing, the request will not incur acceleration fees as the linkage was not accelerated.
When using a global acceleration domain name, you can specify the HTTP or HTTPS transfer protocol. However, if the request is transmitted via a private network Direct Connect line, COS will choose to use the HTTPS protocol to guarantee data transfer security.

Billing Example

Uploading data or accessing a bucket by using a global acceleration endpoint domain name will incur fees calculated by the day. For more information, see Billing Overview and Product Pricing. The following example compares billing between a global acceleration endpoint domain name and a default endpoint domain name:
Business scenario 1
A user uses COS mainly to upload video files which require a high transmission success rate. If he/she uploads 1 GB of video data daily from his/her offices in Xinjiang and Singapore to his/her bucket in Guangzhou, then 30-day fees will be charged as follows:
Upload traffic fees for using an acceleration endpoint domain name: 30 x 1 GB x (0.07 USD/GB + 0.18 USD/GB) = 7.5 USD
Upload traffic fees for using a default bucket endpoint domain name: 30 x 1 GB x （0 USD/GB） = 0 USD
Note:
Upload traffic fees within the Chinese mainland are charged at 0.07 USD/GB, while those outside the Chinese mainland are charged at 0.18 USD/GB when an acceleration endpoint domain name is used. Uploads with a default bucket endpoint domain name do not incur fees.
Business scenario 2
A user uses COS mainly to download video files that require a high transmission success rate. If he/she downloads 1 GB of video data daily from his/her office in Singapore to his/her bucket in Guangzhou, then 30-day fees will be charged as follows:
Download traffic fees for using an acceleration domain name: 30 x 1 GB x 0.18 USD/GB = 5.4 USD
Public downstream traffic fees for using an acceleration domain name: 30 x 1 GB x 0.1 USD/GB = 3 USD
To sum up, the total download traffic fee is 8.4 USD (5.4 + 3 = 8.4).
Note:
The unit price for cross-border download acceleration is 0.18 USD/GB. If you use a global acceleration domain name to download files, public downstream traffic fees and global acceleration downstream traffic fees will be charged.

Private Network Global Acceleration

Last updated：2024-03-25 15:33:39

By leveraging Tencent Cloud's global traffic scheduling capabilities, COS's private network global acceleration routes and resolves user requests intelligently to select the optimal network access linkage, helping you quickly access resources across regions over the private network. It allows you to fully enjoy the stable transfer quality of Tencent Cloud's private network backbone lines, so as to improve the stability and performance of cross-region data transfer.
Many businesses need to pull data across regions over the private network. In the container image distribution scenario, for example, image repositories are usually stored in a region centrally, and the container cluster may be deployed in different regions based on the business needs. During application deployment, the container cluster needs to pull the image from another region, which generates many cross-region requests. In this scenario, a private network global acceleration endpoint can be used to offer a stable private network cross-region Direct Connect line to improve the image distribution stability.
Note:
Using this feature incurs fees as request data transfers are accelerated via Direct Connect lines in the Tencent Cloud private network. For more pricing information, see Pricing | Cloud Object Storage.

How to Use

You can enable global acceleration in the COS console or through APIs.

Using COS console

You can enable global acceleration for your buckets in the COS console. For more information, see Enabling Global Acceleration in the console documentation.

Using RESTful APIs

You can directly use the following APIs to enable global acceleration:
PUT Bucket Accelerate
GET Bucket Accelerate

Access Endpoint Domain Names

After enabling the global acceleration feature, you can access COS files in another region over the private network through a private network global acceleration endpoint in the format of <BucketName-APPID>.cos-internal.accelerate.tencentcos.cn.
For example, a business' application image is stored in bucket examplebucket-1250000000 in Guangzhou region, and its container clusters are deployed in Guangzhou, Beijing, and Shanghai regions. To accelerate container image distribution, the business team enables global acceleration for bucket examplebucket-1250000000, so container clusters in Beijing, Shanghai, and Guangzhou regions can pull the image package over the private network through the private network global acceleration endpoint examplebucket-1250000000.cos-internal.accelerate.tencentcos.cn.
When the container cluster in Guangzhou region pulls the file through the endpoint, COS will intelligently resolve the request to the private network access layer in Guangzhou region to directly pull the file from the storage cluster in the same region.
When the container cluster in Beijing/Shanghai region pulls the file through the endpoint, COS will intelligently resolve the request to the private network access layer in Beijing/Shanghai region and use the private network access layer device to pull the file from the storage cluster in Guangzhou region over the cross-region backbone network Direct Connect line.
Note:
If you use global acceleration, fees will be incurred. Therefore, we recommend that you carefully evaluate whether to use this feature with your actual business needs in mind:
The private network global acceleration endpoint can be used in the Tencent Cloud private network environment only. If your request source is not in the Tencent Cloud private network environment, it cannot be connected to. In this case, you can consider using the default endpoint or the default global acceleration endpoint.
When you use a private network global acceleration endpoint in another Tencent Cloud product such as CVM, TKE, or SCF, the product must be in a region where COS is available; otherwise, it cannot be used. For more information on regions where COS is available, see Regions and Access Endpoints.

Limits

We have outlined below some important factors to note when using a global acceleration endpoint domain name:
The global acceleration domain name will take effect 15 minutes after being enabled.
After the global acceleration domain name is enabled, the maximum bandwidth for a bucket will be allocated based on the business volume of the entire network.
After the global acceleration domain name is enabled, only requests using that domain name will be accelerated. However, the default bucket domain name can still be used.
When using a global acceleration domain name, fees will be incurred only for requests for which linkage is accelerated. For example, if you use a global acceleration domain name to upload data from Beijing to a bucket in Beijing, the request will not incur acceleration fees as the linkage was not accelerated.
When using a global acceleration domain name, you can specify the HTTP or HTTPS transfer protocol. However, if the request is transmitted via a private network Direct Connect line, COS will choose to use HTTPS protocol to guarantee data transfer security.

Monitoring and Alarms

Monitoring and Alarms

Last updated：2025-03-03 19:53:58

Overview

COS statistics such as read and write requests and traffic are collected and displayed based on Cloud Monitor (CM). You can view detailed monitoring data of COS in the COS or CM console.
Note:
This document describes how to get statistics in the COS console. You can call CM APIs to get more detailed data. For more information, see the CM documentation.
Currently, all the metrics reported to CM support all COS regions. For more information, see Regions and Access Endpoints.

Basic Features

CM provides the following modules for COS to implement monitoring and alarming.
Module
Capability
Main Feature
Monitoring overview
Displays the current status of the product
Provides general overview, alarm overview, and overall monitoring information
Alarm management
Supports alarm management and configuration
Supports creating new COS alarm policies, custom messages, and trigger templates
Monitoring platform
Monitors traffic and displays data of user-defined monitoring metrics
Displays your overall bandwidth information and allows you to customize monitoring metrics and data to be reported
Cloud product monitoring
Displays the COS bucket monitoring view
Allows you to query the current monitoring views and data such as read/write requests and traffic for each bucket

Use Cases

Daily management: You can log in to the CM console to view the running status of COS in real time.
Troubleshooting: You can receive alarm notifications when the data of a monitoring metric reaches the threshold. It allows you to quickly notify the exceptions, find out the causes, and fix the issues.

Setting and Querying via Console

You can create an alarm policy for COS in the CM console. If the data of a monitoring metric reaches the specified threshold, you will receive an alarm notification. For detailed directions, see Setting Alarm Policies.
You can go to Cloud Product Monitoring > Cloud Object Storage to view the COS monitoring data (including the monitoring data of all buckets, health status, number of alarm policies, and more). Alternatively, you can go to the COS console to view the data. For detailed directions, see Viewing Statistics and Querying Monitoring Data.

Querying Monitoring Data via APIs

You can call the corresponding Cloud Product Monitoring APIs to view the COS monitoring data. The monitoring metrics are described below. For more information on monitoring APIs, see COS .

Monitoring Metrics

COS Monitoring Metrics

For more information on COS monitoring APIs, see COS .
Note:
COS uses a generic region. Therefore, select Guangzhou for Region when you pull COS monitoring metric data, regardless of where the bucket resides.
When pulling data by using API Explorer, select ap-guangzhou for the Region field.
When pulling data by using an SDK, enter "ap-guangzhou" for the Region field.
The statistical granularity (period) may vary by metric. You can call the DescribeBaseMetrics API to obtain the period supported by each metric.
For more information on 3xx, 4xx, and 5xx status codes, see Error Codes.

Dimensions and Parameters

Parameter
Dimension
Description
Format
&Instances.N.Dimensions.0.Name
appid
Dimension name of the root account APPID
Enter a string-type dimension name: appid
&Instances.N.Dimensions.0.Value
appid
Specific root account APPID
Enter a root account APPID, such as 1250000000
&Instances.N.Dimensions.1.Name
bucket
Dimension name of the bucket
Enter a string-type dimension name: bucket
&Instances.N.Dimensions.1.Value
bucket
Specific bucket name
Enter a specific bucket name, such as examplebucket-1250000000

Input Parameters

To query COS monitoring data, the values of the input parameters are as follows:
&Namespace=QCE/COS &Instances.N.Dimensions.0.Name=appid &Instances.N.Dimensions.0.Value=root account APPID &Instances.N.Dimensions.1.Name=bucket &Instances.N.Dimensions.1.Value=bucket name

Monitoring Description

Monitoring interval: CM supports multiple monitoring intervals, including real time, last 24 hours, last 7 days, and user-specified period, with time granularities of 1 minute, 5 minutes, 1 hour, and 1 day.
Data storage: 1-minute monitoring data can be stored for 15 days, 5-minute data for 31 days, 1-hour data for 93 days, and 1-day data for 186 days.
Alarm display: CM integrates the monitoring data of COS and displays the data in graphs. Alarm notifications can be sent to you according to the predefined alarm metrics of your product. In this way, you can stay informed of the overall running status.
Alarm settings: You can set the threshold for the monitoring metrics. When the monitoring data meets the alarm condition that is set, CM will send the alarm notifications to the specified users. For more information, see Alarm Overview and Setting Alarm Policies.

Monitoring Reports

Viewing Data Overview

Last updated：2025-05-28 17:18:27

Overview

You can go to the Statistic page in the COS console to view the number of buckets/objects/requests, storage usage, traffic, and other data.
Note:
If a sub-account needs to view the statistics, it needs to have permission to access the bucket list. For more information, see Adding a Preset Policy in Accessing Bucket List Using Sub-Account.
Except for the number of buckets, other data is not real-time and has a delay of about 2 hours. The data is for monitoring purposes only. To view the accurate billing data, go to Billing Center.

Viewing Line Charts

Directions

1. Log in to the COS console and click Statistic > Basic Statistic on the left sidebar.
2. On the Basic Statistic page, you can view the following information:



Number of Buckets: number of buckets that have been created.
Total Objects: number of objects that have been created in all buckets. Supports drop-down selection to view the number of objects for different storage types. Click View All in the upper right corner to quickly view detailed data for each storage type and support data export. For details, see Viewing and Exporting Overview Data.
Average Daily Usage for This Month: The average daily usage by storage class (STANDARD, MAZ_STANDARD, STANDARD_IA, MAZ_STANDARD_IA, INTELLIGENT_TIERING, MAZ_INTELLIGENT_TIERING, ARCHIVE, MAZ_ARCHIVE and DEEP_ARCHIVE). Daily storage usage = sum of the "5-minute usage"/288 (number of statistical points). Average daily storage usage for this month = sum of the daily storage usage in this month/number of days in this month. Click View All to quickly view detailed data for each storage type and support data export. For details, see Viewing and Exporting Overview Data.
Storage Usage: Storage usage for STANDARD, MAZ_STANDARD, STANDARD_IA, MAZ_STANDARD_IA, INTELLIGENT_TIERING, MAZ_INTELLIGENT_TIERING, ARCHIVE, and DEEP_ARCHIVE. You can view the data for the past 90 days, and the time range for each query cannot exceed 31 days.
Traffic: Public/private downstream traffic,CDN origin-pull traffic and Global acceleration downstream traffic for STANDARD, MAZ_STANDARD, STANDARD_IA, MAZ_STANDARD_IA, INTELLIGENT_TIERING, and MAZ_INTELLIGENT_TIERING. You can view the data for the past 90 days, and the time range for each query cannot exceed 31 days.
Note:
Global acceleration downstream traffic is case-insensitive to storage type.
Request Count: Number of read/write requests for STANDARD, MAZ_STANDARD, STANDARD_IA, MAZ_STANDARD_IA, INTELLIGENT_TIERING, and MAZ_INTELLIGENT_TIERING. You can view the data for the past 90 days, and the time range for each query cannot exceed 31 days.
Request Success Rate: Request success rate for STANDARD and MAZ_STANDARD. You can view the data for the past 90 days, and the time range for each query cannot exceed 31 days.
Data Retrieval: The amount of data retrieved for STANDARD_IA, MAZ_STANDARD_IA, INTELLIGENT_TIERING, MAZ_INTELLIGENT_TIERING, ARCHIVE, and DEEP_ARCHIVE. For ARCHIVE, the amount is classified into expedited retrievals, standard retrievals, and bulk retrievals. You can view the data for the past 90 days, and the time range for each query cannot exceed 31 days.
3. You can click the download button in the upper-right corner of each chart to download the data.

View and Export Overview Data

Operation Steps

You can quickly view and export detailed data for each storage type of Total Objects and Average Daily STANDARD Usage for This Month in the following ways.
1. In the overview module on the Basic Statistic page, click View All in the upper right corner, and you will see the overview data .
2. Click Export data to download the overview data.




Viewing Bucket-Level Data

Directions

1. Log in to the COS console, click Bucket List on the left sidebar, and click Statistical Data.
2. On the Statistical Data page, you can view the bucket data such as storage usage, data retrievals, traffic, and request count within a specified period of time.

Querying Monitoring Data

Last updated：2025-11-14 17:11:20

Overview

COS allows you to monitor your stored data. The COS data monitoring window displays information regarding request quantity, traffic, return code, data retrieval, etc. You can also query the details and trends of data in different storage classes for different time periods. The following describes how to view the monitoring data of a single bucket with a root account or sub-account.
Note:
To view the summary of all data stored under your account, go to Data Monitoring in the COS console.

Querying with Root Account

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Find the target bucket and click Monitor on the right. You can also select the target bucket and click Data Monitoring on the left sidebar on the bucket details page.
4. On the monitoring data page, view the information.

Monitoring items are as described below:
Storage: Storage usage of each storage class.
Number of Objects: Number of objects (including incomplete multipart uploads) stored in the bucket.
Number of Incomplete Multipart Uploads: Number of incomplete multipart uploads stored in the bucket. If an ongoing upload is suspended or canceled, the corresponding files will be stored in the bucket as incomplete multipart uploads.
Note:
To query the number of objects in a folder, see Viewing Folder Details.
If versioning is enabled, multiple versions of an object are counted as separate objects.
Requests: Number of all requests (including GET and PUT requests), as well as read/write requests of STANDARD and STANDARD_IA.
Traffic: Public/Private network traffic, CDN origin-pull/cross-region replication traffic, and total upload traffic over public and private networks.
Return Code: Number of 2xx, 3xx, 4xx, and 5xx status codes as well as their proportion.
Data Retrieval: Statistics of data retrievals from STANDARD_IA.
Note:
Under Details in the Current section, you can query monitoring data (including storage usage, number of objects, requests, traffic, return code, and data retrieval) of different time periods, such as today, yesterday, last 7 days, last 30 days, or a custom time period.
In the This month section, you can view data for the month, including the daily average storage usage of each storage class as well as total traffic (total public network traffic, total CDN origin-pull traffic, and total cross-region replication traffic).
Storage and Number of Objects show only data after March 1, 2020. For more detailed data, go to Billing Center, select a time period, and export the data.

Querying with Sub-account

To query monitoring data with a sub-account in the console, you need to first grant the sub-account relevant permissions.
You can grant such permissions by using a policy template or custom access policy.

Authorizing by policy template

1. Log in to the CAM console as the root account and select Users > User List to enter the user list page.
2. Find the target sub-account and click Authorize in the Operation column on the right.
3. Search for and select the QcloudMonitorFullAccess policy in the pop-up window and click OK to associate it with the sub-account. Then, the sub-account can access monitoring reports.
Note:
This policy template grants the sub-account the full access to CM. To protect the security of your account, you can customize an access policy to grant only read permissions to the sub-account.

Authorizing by custom access policy

1. Log in to the CAM console as the root account.
2. On the left sidebar, click Policies > Create Custom Policy > Create by Policy Syntax.
3. Select Blank Template and click Next.
4. Copy and paste the following policy syntax into the Policy Content input box. You can rename the policy as needed. Policy syntax:
{
 "version": "2.0",
 "statement": [
     {
         "effect": "allow",
         "action": [
             "monitor:GetMonitorData"
         ],
         "resource": "*"
     }
 ]
}



5. Click OK. After the policy is created successfully, you can associate it with the sub-account as instructed in Authorizing by policy template.

Setting Alarm Policies

Last updated：2024-01-06 15:14:56

Overview

You can leverage the alarm policy feature of Cloud Monitor to set threshold-reaching alarms for COS monitoring metrics. An alarm policy must include the policy name, policy type, trigger condition, alarm object, and alarm notification template. You can create an alarm policy for COS as instructed below.
Note:
Tencent Cloud’s Cloud Monitor enables users to monitor cloud resources in real time and provides alarm services. Users can set alarm policies for COS monitoring metrics to query the alarm history and receive alarm notifications. For more information, please see Creating Alarm Policy.

Directions

1. Log in to the COS console. On the Overview page, click Configure Alarm Policy in the Alarm Configuration section.
Note:
Alarm Configuration can also be found on the Overview page of each bucket.
2. Configure a new alarm policy as shown below:
Configuration Type
Configuration Item
Description
Basic info
Policy name
A custom policy name
Remarks
Remarks for the policy
Monitor Type
Choose Cloud Product Monitoring.
Policy type
Choose COS.
Project
Setting the policy project allows you to:
Manage alarm policies. Alarm policies of a project can be quickly located in the alarm policy list.
Manage instances. You can choose the project as needed. Instances of the project can be quickly located in Alarm Object. You can distribute Tencent Cloud services to the desired project according to your business types. After the project is created, you can distribute resources to projects in the console of each Tencent Cloud service. Some Tencent Cloud services cannot be distributed to a project. If you do not have project permission, please see Cloud Access Management (CAM) for authorization.
Configure alarm rule
Alarm object
If you choose Instance ID in the drop-down list, select the bucket you want to add an alarm to.
If you choose Instance Group in the drop-down list, the alarm policy is bound to the selected instance group. If there is no instance group, you can click Create Instance Group on the right to create a group for the bucket first.
If you choose All Objects in the drop-down list, the alarm policy is bound to all buckets the current account has permission on.
Select template
Select a configured template from the drop-down list. For more information on the configuration, see Configuring Trigger Condition Template. If the newly created template is not displayed, click Refresh on the right.
Manual configuration(Metric alarm)
Trigger condition: Consists of metric, comparison, threshold, statistical period, and the number of consecutive periods.For example, if the metric is set to `STANDARD storage read requests`, comparison to `>`, threshold to `80` times, statistical period to `5` minutes, and the number of consecutive periods to `Last 2 periods`, STANDARD storage read requests are collected every 5 minutes, and if the number of STANDARD storage read requests in a bucket is greater than 80 in 2 consecutive periods, the alarm will be triggered.
Alarm frequency: You can set a repeated notification policy for each alarm rule. In this way, an alarm notification will be sent repeatedly at a specified frequency when an alarm is triggered.

Frequency options: do not repeat, once every 5 minutes, once every 10 minutes, at an exponentially increasing interval, and other frequency options.

An exponentially increasing interval means that a notification is sent when an alarm is triggered the 1st time, 2nd time, 4th time, 8th time, and so on. In other words, the alarm notification will be sent less and less frequently as time goes on to reduce the disturbance caused by repeated notifications.

Default logic for repeated alarm notifications: The alarm notification will be sent to you at the configured frequency within 24 hours after the alarm is triggered. After 24 hours, the alarm notification will be sent once a day by default.

Configure alarm notification
Alarm notification
Select a preset or custom notification template. Each alarm policy can be bound to three notification templates at most.
Advanced configuration
Auto scaling
If enabled, the auto scaling policy can be triggered when the alarm condition is met.
3. After configuring the above information, click Complete.

Operations Center

Security Management

Last updated：2025-11-18 15:47:53

Introduction

Cloud Object Storage (COS) console provides a security management feature to help users identify, assess and control bucket configuration risks, thereby ensuring data security and compliance. Through risk detection, evaluation reports and control measures, it can effectively reduce potential security risks caused by improper bucket configurations, such as data leakage and traffic theft.
Note:
COS provides security management service based on Tencent Cloud Smart Advisor (TSA). For first-time use, activate the cloud advisor service and authorize the service role first.
To ensure data security, the sub-account must have bucket list access permissions to use security management normally. If there is no permission to view, contact the root account to assign the corresponding permission. For details, see add preset policy QcloudCOSGetServiceAccess.
Security management feature can be used for free.

Initiating Assessment

1. Log in to the COS Console.
2. In the left navigation bar, select Ops Center > Security Management to access the Security Management page.
3. In the overview area, click reassess to display security assessment in progress, which means initiating an assessment task.
Note:
After first-time activation, the system will initiate an assessment task by default. Besides, a daily assessment task is automatically initiated in the morning. For details, see Latest Evaluation Time.
Note: After the assessment task is initiated, it takes some time. Wait patiently. You may close the page after initiating the assessment task and enter the page for viewing the assessment result later.


Risk Overview

The Risk Overview module mainly shows the overall risk situation of bucket configurations. This includes the number of evaluation items under three levels: high-risk items, mid-risk items, and Healthy items. It also supports viewing week-on-week comparisons for each risk level. Directions for viewing the assessment overview:
1. Access the Security Management page. After task completion, view the overall risk situation on the right side of the Overview section. The assessment result is based on the latest evaluation time statistics. If no manual initiation is performed on the current day, the result of the automatic assessment from that morning will be shown. For more details, please refer to Initiating Assessment.

2. Click a certain risk level area to automatically filter the risk assessment results under the current risk level.


Risk Management

Risk management mainly includes five modules: risk assessment result, result detail and manage, report download, assessment item settings, and historical risk, making it convenient for users to evaluate and fix detected risk configuration items.

Risk Assessment Result

Risk assessment results are mainly divided into five categories based on bucket configuration attributes: hotlinking, security, reliability, performance, and cost. You can intuitively view assessment results under different categories, including assessment item name, description, and the number of risk buckets under each item. Directions for viewing risk assessment results:
1. Access the Security Management page and view the assessment result section.
2. By default, all assessment results are displayed, i.e., the evaluation results of all detected bucket configuration items. Three are shown by default. Click expand more to view other assessment results.
3. To filter all assessment results, click on the right filtering box to perform dropdown selection.

4. You can also click to switch at the top Tab bar to view assessment results under different categories, such as results under the hotlinking category.


Result Details and Manage

In the assessment result section, you can view the bucket list with existing risks under each assessment item, learn about the trigger rule, and additionally obtain appropriate optimization suggestions for quick management. Directions:
1. Access the Security Management page and view the assessment result section.
2. Click a certain assessment item card, and the result details and management page will automatically pop up on the right.
3. The assessment item name, assessment item description, number of buckets at risk, and trigger rule, as well as optimization suggestions, are shown at the top of the page. The list of buckets under evaluation and ignored is displayed at the bottom.
4. You can operate on the designated bucket as per the optimization suggestion. Click the bucket in the list to quickly go to configuration.

5. Under the buckets under evaluation list, you can ignore the bucket by clicking on the right side of the list. The bucket will be added to the ignored buckets list and will not be detected during evaluation.

6. Under the ignored buckets list, you can unignore the bucket by clicking on the right side of the list. The bucket will be added to the buckets under evaluation list and will continue detection during evaluation.


Download Reports

You can download the security reports for all assessment items and each assessment item. Directions:
Note:
This report only shows assessment items with existing risks and detailed information of related risk buckets. Risk-free items will not be reflected in the report.

All Assessment Items

1. Access the Security Management page and view the overview section.
2. Click Download Report to download and view the security report for all assessment items.


Single Assessment Item

1. Access the Security Management page and view the assessment result section.
2. Locate the assessment item card you need to download, click the download icon to download and view its security report.


Assessment Item Settings

You can enable or disable an assessment item and adjust its rule as needed. Directions:

Enable or Disable the Assessment Item

1. Access the Security Management page and view the assessment result section.
2. Click Assessment Item List, navigate to Tencent Cloud Smart Advisor and turn on or off the assessment item switch. For more details, please refer to Enable or Disable Inspection Assessment Items.
Note:
Note: If you disable an assessment item in the Tencent Cloud Smart Advisor console, the inspection item will no longer be included in the next risk assessment and will show as a risk-free bucket by default.


Adjusting Assessment Item Rules

1. Access the Security Management page and view the assessment result section.
2. Click Assessment Item List, navigate to Tencent Cloud Smart Advisor and click Modify Rules. For more details, please refer to Adjusting Inspection Assessment Item Custom Rules.
3. At the same time, you can also click the settings icon on a certain assessment item card to quickly redirect to Tencent Cloud Smart Advisor and perform rule adjustment.
Note:
Quick redirect link is applicable only to assessment items that support rule adjustment.


Historical Risk

You can query the risk trend change and view the historical risk details. The operation details are as follows.
1. Access the Security Management page and view the assessment result section.
2. Click Historical Risk, navigate to TSA to view risk trend changes and historical risk details. You can choose risk level and time range as needed to retrieve more detailed trend information. For more details, please refer to Query Risk Trend and Query Historical Risk Details.


Fraudulent Ban Bucket Query

COS supports the traffic anti-fraud feature. The system will intelligently detect abnormal traffic, automatically block abnormal buckets, and send notifications. If a bucket is banned due to traffic theft, you can view the blocking record details in this region, including the bucket name, region, and blocking time.

Bucket name: Show banned bucket names.
Region: Show the region where the bucket is located.
Blocking time: Show the specific time when the bucket was banned.
Note:
To avoid high costs caused by abnormal traffic before automatic blocking, take security measures as needed. For more information, check anti-fraud guidelines.
If the traffic growth is expected, contact us to apply for unblocking.

Data Processing

Image Processing

Basic Image Processing

Last updated：2024-01-06 15:14:56

Overview

This document describes how to use COS Image Processing in the console, with the following two methods available for you. For more information on image processing, see Image Processing Overview.
Adding parameters to an image URL: You can process an image by adding parameters after the object URL of the image.
Using an image style: You can save different processing results by creating styles, and use such styles to standardize your image processing. The style created is an assembled template designed to process parameters in real time when an image is downloaded.
Note:
Image Processing is available only in public cloud regions.
Image Processing is charged by CI. For detailed pricing, see Basic image processing fee in Billing and Pricing.

Adding URL parameters

1. Log in to the COS Console.
2. Locate the bucket that stores the image, and click the bucket name to enter the bucket management page.
3. Click Details in the Operation column for the image to enter its details page.
4. Copy the Object URL and paste it into the address bar of your browser.
Note:
To process an image, you need the write permission on objects. For object permissions setting, see Setting Object Access Permission.
5. In your address bar, add parameters after the object URL using the format below. For more image processing parameters and instructions, see the Cloud Infinite API documentation Basic Image Processing.
Object URL?processing API name/processing operation name/processing parameter
Note:
If the access permission on the image file is private-read, you need to add image processing parameters to the signed address.
Example: scaling an image to 50% of its original size Suppose that the original image is displayed as below, the access permissions on the object are public read and private write, and the object URL is https://examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/sample.jpeg.
img

Next, add the following parameters to the URL:
Scaling API: imageMogr2
Scaling operation name: thumbnail
Processing parameter: !50p
Now, you obtain a new URL with added parameters https://examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/sample.jpeg?imageMogr2/thumbnail/!50p. Paste this URL to your address bar, press Enter, and you will see a scaled image as shown below:


Using an image style

Image Style can help you present different processing parameters in the format of template to standardize your image processing. The style created is an assembled template designed to process parameters in real time when an image is downloaded. Now, let’s take Image Style: width: 480px, height: 270px as an example, and introduce how to use this feature:
1. Log in to the COS console.
2. Locate the bucket that stores the image, and click the bucket name to enter the bucket management page.
3. In the left menu bar, click Data Processing > Image Processing to enter the Style Management section. Separator: Style separators are symbols that separate filenames and processing styles, including -, _, /, and !. Select ! here and save.
4. Click Add Styles and set fields as follows:
Style Name: Enter your custom style name, such as “yunstyle”.
Note:
Note that style names are case-sensitive and cannot be modified once saved.
For the purpose of clarity, the separator you have enabled cannot appear in the style name.
Editing Mode: Select Basic.
Resize Mode: Select Scale-only.
Scaling: Select Fixed height and width.
Size: width: 480px, height: 270px.
Progressive Display: Once enabled, the images you access will be displayed progressively. It is disabled by default. Retain the default setting here.
Output: Choose a format of the output image. Retain the default value Original here.
5. Once your configuration is done, You can click the Preview button on the right for preview.
6. After the preview, click Save, and you will see that an image style named “yunstyle” has been added.
Note:
Up to 100 styles can be set for a single bucket.
The settings take effect in 30 minutes on average.
Changing a separator requires clearing the cache. It takes at least 24 hours for separator changes to take effect globally.
Canceling a separator used may cause product feature malfunctions.
For more information on image styles, see Setting Styles.
7. Go to the object details page, copy the object URL, and enter your separator and style name after the URL in the following format:
Object URL + Separator + Processing style name
Now, you can obtain a final object URL https://examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/sample.jpeg!yunstyle. Paste it to your address bar, press Enter, and you will see a scaled image shown as below:

If you need to set a signature for the styled image, change the URL to https://examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/sample.jpeg!yunstyle?q-sign-algorithm=<signature>, and use /sample.jpg!yunstyle as the resource to calculate the signature <signature>.

Setting Image Advanced Compression

Last updated：2024-01-06 15:14:56

Overview

Image Advanced Compression is a COS feature based on Cloud Infinite. It allows you to easily convert images into formats that provide a high compression ratio, such as TPG and HEIF. This effectively reduces the transmission time, loading time, and the use of bandwidth and traffic.
Note
You can use Image Advanced Compression to convert JPG, PNG, or WebP images into TPG or HEIF.
To use the TPG format, ensure that the environment where images are loaded supports TPG decoding. CI provides the TPG decoder−integrated SDK for iOS, Android clients to facilitate quick integration with TPG.
Currently, iOS 11 or later and Android P have native support for the HEIF format.
Image Advanced Compression is charged by CI. For detailed pricing, please see Billing and Pricing.

Directions

1. Log in to the COS console.
2. Click Bucket List in the left sidebar.
3. Click the bucket you want to enable Image Advanced Compression for.
4. Click Data Processing > Image Processing.
5. Find the Image Advanced Compression area, click Edit, enable the status, and click Save.

Once Image Advanced Compression is enabled, you can call the Image Advanced Compression API to convert images in the bucket into TPG/HEIF upon download.

File Processing

Setting File Processing

Last updated：2024-01-06 15:14:56

Overview

The COS console allows you to preview documents in a bucket. This document describes how to use the document preview feature in the console. For information on document preview, see Document Preview Overview.
Note:
The document preview feature is available in public cloud regions in the Chinese mainland and Silicon Valley, Virginia, Frankfurt, and Singapore. Currently, only document-to-image conversion preview is supported in Singapore and Silicon Valley.
Document preview is charged by CI. For detailed pricing, see Billing and Pricing.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar, and then click the bucket for which you want to enable the document preview feature.
3. Click Data Processing > Document Processing. In the Document Preview area, click Edit and toggle on the status
4. To preview a document, add parameters to the document URL. The parameters are described as follows:
<BucketName-APPID>.cos.<Region>.myqcloud.com/<objectkey>?ci-process=doc-preview&page=<page>&srcType=<srcType>
objectkey: Object key, which can be understood as a file path.
ci-process: CI's processing capability, which is fixed at doc-preview for document preview.
page: Number of the page to be converted, which is counted from 1.
srcType: Source data type. Currently, the file conversion feature determines the source data type according to the file extension of the COS object. If the object has no extension, you can set this value.
Note:
If you use the URL mode, only a single-page document or the first page of a multi-page document can be processed. If you want to preview more content, use Document Preview API.

Media Processing

Enabling Media Processing

Last updated：2024-01-06 15:14:56

Overview

This document describes how to enable the media processing feature in the COS console. After enabling, you can create workflows or tasks to perform media processing operations. For more information, see Media Processing Overview.
Note:
To use the media processing feature, first enable Cloud Infinite (CI). Media processing fees incurred will be charged by CI. For detailed pricing, see CI Media Processing Fees.

Directions

1. Log in to the COS console.
2. Choose Bucket List on the left sidebar.
3. Click the name of the bucket that you want to operate.
4. On the left sidebar, click Data Processing > Media Processing.
5. Click Edit and set the status to enabled.
6. Click Save.

Subsequent Operations

After enabling media processing, you can perform the following operations:
For the specified files in a bucket: You can create a media processing task to perform media processing on them. For details, see Creating a Task.
For incremental files in a bucket: You can create a workflow to automatically process and save the files after they are uploaded to the bucket. For details, see Creating a Workflow.

File Processing

Configure File Hash Task

Last updated：2025-05-27 10:50:16

Overview

For existing files in a bucket, you can create a task to perform hash calculation on the files.
Note:
Hash calculation tasks only support calculating files less than 50 GB.
After creating a task, appropriate service fees will occur, which are collected by Cloud Infinite. For detailed billing information, see File Processing Fees.
Supported regions: Beijing, Nanjing, Shanghai, Guangzhou, Chengdu, Chongqing.

Creating a Hash Calculation Task

File hash calculation provides computational operations for multiple hash algorithms and supports automatically adding to the custom header attribute of a file upon completion of the hash calculation.

Operation Steps

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to enter the bucket list management page.
3. Find the bucket that requires operation, click the Bucket Name, and go to the bucket management page.
4. In the left sidebar, select Task and Workflow > Task Management to enter the task management page.
5. Select the File Processing tab, choose the Hash Calculation task type, click Create Job, and configure according to the following information:



Source File URL: Enter or select the path where the file to be processed is located. Note that the file path cannot start or end with /.
Hash Algorithm: Supports three hash algorithms: MD5, SHA1, and SHA256.
Custom Header: You can choose to automatically add it to the custom header after the hash calculation is complete.
Callback URL: The callback address information bound to the queue. For details, see Queues and Callbacks.

Viewing Tasks

You can view all tasks of different task types within a specified period on the task page. You can click Task Status to filter and view tasks in different statuses. You can also search for a specified task by task ID in the search box.

Operation Steps

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to enter the bucket list management page.
3. Find the bucket that requires operation, click the Bucket Name, and go to the bucket management page.
4. In the left sidebar, select Task and Workflow, click Task to enter the task management page.
5. Click View on the right of a task to view its detailed information.
Note:
Task includes 6 statuses: execution successful, execution failed, executing, unexecuted, suspended, and canceled.
Only support querying task records of the last month.

Configure File Decompression Task

Last updated：2025-05-21 18:53:11

Overview

For the compressed package file in the existing bucket, you can create a task to perform the decompression operation.
Note:
Currently supports decompression formats: zip, tar, gz, 7zip, rar, apk.
Decompression tasks support decompressing files less than 5 TB.
After creating a task, appropriate service fees will be incurred and charged by the Cloud Infinite (CI) service. For detailed billing information, see File Processing Fee.
Supported regions: Beijing, Shanghai, Guangzhou, Chengdu, Nanjing, Chongqing, Hong Kong (China), Singapore, Silicon Valley.

Create File Decompression Task

The file decompression feature supports decompression operations for various compressed archive formats.

Operation Steps

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to enter the bucket list management page.
3. Find the bucket that requires operation, click the Bucket Name, and go to the bucket management page.
4. In the left sidebar, select Task and Workflow, click Task Management, and enter the task management page.
5. Select the File Processing tab, choose the File Decompression task type, click Create Job, and configure according to the following information:



Source File URL: Enter or select the path where the file to be processed is located. Note that the file path cannot start or end with /.
Compression key: You can enter a password to decompress an encrypted compression package. Currently only supports decryption of zip, 7zip, and rar compression formats.
Destination Bucket: The bucket where files are saved after decompression. Only buckets that have enabled the file processing service in the same region are supported.
Destination Path: The specific path where files are saved after decompression. If not specified, files are saved in the root path of the target bucket by default.
Keep compressed package name: Specify the directory of files after extraction. Whether to retain the name of the archive. The following options are available:
No: After file decompression, files will be saved directly in the destination path (the name of the archive will not be retained).
Yes: On the basis of the destination path, after file decompression, files will be saved in a directory named after the archive name.
Retain the full path of the compressed package: On the basis of the destination path, after file decompression, files will be saved in a directory named after the full path of the compressed package (the directory includes the name of the compressed package).
Callback URL: The callback address information bound to the queue. For more details, see Queues and Callbacks.

Viewing Tasks

You can view all tasks of different types within the specified period on the task page. You can click task status to filter and view tasks in different states. You can also use the search bar to find a specified task by its task ID.

Operation Steps

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to enter the bucket list management page.
3. Find the bucket that requires operation, click the Bucket Name, and go to the bucket management page.
4. In the left sidebar, select Task and Workflow, click Task, and enter the task management page.
5. Click View on the right of a task to view its detailed information.
Note:
Task includes 6 statuses: execution successful, execution failed, executing, unexecuted, paused, canceled.
Only support querying task records in the last month.

Configure Multi-File Compression Task

Last updated：2025-05-21 18:53:12

Overview

Cloud Object Storage (COS) supports file zipping functionality for multiple files in the cloud. It supports batch packaging of files in a certain path on COS or packaging based on several file single links provided by you.
Note:
Use the packaging method for multiple files. It supports up to 1,000 files with a total size of no more than 50 TB.
Use the packaging method with directory or manifest, supporting a maximum of 10 million files with a total size not exceeding 50 TB.
After creating a task, appropriate service fees will occur, which will be collected by Cloud Infinite. For detailed billing information, see File Processing Fees.
Supported regions: Beijing, Shanghai, Guangzhou, Chengdu, Nanjing, Chongqing, Hong Kong (China), Singapore, Silicon Valley.

Create a Multi - File Zipping Task

Operation Steps

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to enter the bucket list management page.
3. Find the bucket that requires operation. Click the Bucket Name, and go to the bucket management page.
4. In the left sidebar, click Task and Workflow > Task Management to enter the job management page.
5. Select the File Processing tab, choose the Multi-file Compression task type, click Create Task, and configure according to the following information:



Select source file: The following three methods are supported for file packaging and compression:
File: Manually select multiple files in the bucket and pack and compress them;
Select a directory in the bucket manually and pack and compress the entire directory.
URL list file: Organize the list of files to be packaged into a CSV manifest file and save the manifest file in a bucket. Select the manifest file in the bucket to package and compress the files in the list.
Directory Flattening: After using directory flattening, the source directory structure will be removed. For example, if the source file URL is https://example.com/source/123/test.mp4, the file path here is source/123/test.mp4.
If flattened is used, the file in the compressed package is test.mp4;
If flattened is not used, the file in the compressed package is source/123/test.mp4.
Note:
If using directory flattening, existing files with the same name under different directories will be automatically overwritten and only one will be kept in the compressed package.
Compression Format: Supports compression into three formats: zip, tar, gz.
Compression method: Support setting the compression effect:
Default: Moderate compression speed and compressed file size.
Prefer higher compression speed: Shorter compression time, but relatively larger compressed file size.
Prefer higher compression quality: Longer compression time, but smaller compressed file size.
Encryption: Encryption is supported during compaction and is only supported when the target format is zip.
Ignore error: You can choose to terminate the packaging operation when encountering issues such as being unable to obtain a single file when packaging.
Destination Bucket: The bucket where files are saved after decompression. Only buckets that have enabled the file processing service in the same region are supported.
Destination Path: The specific path where files are saved after decompression. If not specified, they are saved in the root path of the target bucket by default.
Destination File Name: The file name after packaging and compression.
Callback URL: The callback address information bound to the queue. For details, see Queues and Callbacks.

Viewing Tasks

You can view all tasks of different task types within a specified period on the task page. You can click Task Status to filter and view tasks in different statuses. You can also search for a specified task by task ID in the search box.

Operation Steps

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to enter the bucket list management page.
3. Find the bucket that requires operation. Click the Bucket Name, and go to the bucket management page.
4. In the left sidebar, select Task and Workflow, click Task Management to enter the task management page.
5. Click View on the right of a task to view its detailed information.
Note:
Task includes 6 statuses: execution successful, execution failed, executing, unexecuted, paused, canceled.
Only support querying task records in the last month.

Function Service

Setting CDN Cache Purge

Last updated：2024-10-24 18:44:41

Overview

CDN cache purge is a data purge feature provided by COS based on the SCF service. It can automatically purge the data cached on CDN edge nodes. After you add a trigger rule for a bucket, when a file is updated to the bucket, the SCF function preconfigured by COS will be automatically triggered to purge the cached data.
Note:
If you added a cache purge rule for your bucket in the COS console, the resulting purge function will also appear in the SCF console. Do not delete this function; otherwise, your rule may not take effect.
Regions where SCF is available support CDN cache purge, including Guangzhou, Shanghai, Beijing, Chengdu, Hong Kong (China), Singapore and Silicon Valley. For more supported regions, see Serverless Cloud Function.
CDN cache purge may fail due to factors such as unstable network connection. In these cases, you can click View Log for the function created in the COS console to enter the SCF console and view the error log details for troubleshooting.
The CDN cache purge feature depends on the SCF service, which provides a free tier. Excessive usage will be billed at SCF prices. For more information, see Pricing. For this feature, the more often you purge the cache, the more invocations you will need to use.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar and click the name of the target bucket.
3. Click Function Service > CDN Cache Purge Function on the left sidebar.


Note:
If you haven't activated SCF, go to the SCF console to activate it and authorize the service as instructed.
4. Click Add Function and configure the following in the pop-up window:


Function Name: It uniquely identifies a function and cannot be modified after being set. You can view the function in the SCF console.
Event Type: An event is an operation that triggers SCF. Take upload as an example. You can initiate an upload by calling the PUT Object or Post Object API. If you select PUT as the event type, only uploads through the PUT Object API can trigger the function to purge the cache on CDN edge nodes.
Trigger Condition: You can specify the file source scope of the triggered files, such as all buckets, specified prefix, or specified suffix. After you specify a prefix or suffix, the rule only applies to the matched file sources.
Specified Domain: Specify the CDN domain name to be purged.
SCF Authorization: CDN cache purge requires authorizing SCF to call the CDN API PurgeUrlsCache. In this way, SCF can replace your CDN cache with the latest data CDN pulls from the COS origin.
5. Click Confirm.
Click Log to view the historical operations of CDN cache purge.
Click Delete to delete an unwanted CDN cache purge rule.

Content Moderation

Moderation Details

Last updated：2025-11-18 16:23:19

Overview

The COS content moderation service intelligently moderates the multimedia content of images, videos, audios, documents, and text. It helps you effectively identify non-compliant content such as pornographic, vulgar, illegal, disgusting, and offensive information to avoid operational risks.
After files are moderated successfully, you can view the moderation results by condition and manually process them on the Moderation Details page.

Directions

Filtering results

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the target bucket to go to the configuration page.
4. On the left sidebar, select Sensitive Content Moderation > Moderation Details to enter the Moderation Details page.
5. Select the conditions as needed.
Scope: You can view the results of moderation through API calls, automatic moderation, and historical data moderation.
File type: You can view the moderation results of images, videos, audios, and text.
Level-1 category of moderation result: It refers to the level-1 label of the moderation result. You can specify it to filter the results.
Level-2 category of moderation result: It refers to the level-2 label of the moderation result, which is more specific than the level-1 category. You can specify it to filter the results.
Moderation Result: Moderated files are categorized into three types: sensitive, suspected, and normal files. You can view a type of file or all files.
Sensitive: Images whose score falls within the range of [91,100].
Suspected: Images that are suspected to be sensitive and whose score falls within the range of [61,90]. For such images, the system cannot determine whether they are sensitive, so we recommend you determine it through human moderation.
Normal: Images whose score falls within the range of [0,60].
Block Status: You can view the moderation results of blocked, normal, or all files.
Moderation Status: You can filter the moderation status, including under review, succeeded, and rejected.
Moderation Policy: You can filter moderation results by moderation policy (BizType).
Moderation time: You can view the moderation results in the specified time period.
Note:
If you rename a file or modify its metadata, the file will be considered a newly uploaded file and have a new moderation result.
Image Score: If you select All for Moderation Result, you can filter files by customizing the file moderation score interval.
Object Name: You can enter a filename to view the moderation result of the specified file.
6. Click Query to view the moderation results.
Note:
The Moderation Details page only displays the details of moderations called in the console but not those called through an API or SDK.

Exporting results

After filtering results, click Export to export the results as a .csv or .xlsx file.
The fields in the moderation result file are as detailed below:
Field
Description
Moderation record ID
Unique ID of the moderation job (also known as job ID), through which you can query the moderation result.
Object name
Name of the moderated file.
Note: This field is empty if you use the API to directly moderate text content.
Moderation result type
Moderation result type of the file such as Porn or Ad. If the moderation result is Normal, this field will be empty.
Moderation score
The confidence the moderation result hits the moderation scene. Value range: 0–100. The higher the value, the more likely the content hits the currently returned moderation scene.
For example, 99 means that the content is very likely to be pornographic.
Moderation result subtype
Moderation subtype of the file such as Sexy. If the moderation result is Normal, this field will be empty.
Moderation result
Moderation results include Normal, Sensitive, and Suspected. The result is determined by the moderation score.
Normal: The moderation score is 0–60.
Sensitive: The moderation score is 91–100.
Suspected: The moderation score is 61–90.
Blocked
Whether the moderated file is blocked. You can configure automatic blocking upon moderation in Automatic Moderation and Historical Data Moderation.
File size
Unit: B.
Moderation time
Time when the file is moderated.
File path
Location of the file in the bucket.
Note: This field is empty if you use the API to directly moderate text content or a file not in COS.
Original link
File URL.
Note: This field is empty if you use the API to directly moderate text content.
Moderation method
File moderation method. Valid values:
Notify: Automatic moderation.
Task: Historical data moderation.
API: Moderation through API call.
File type
Data source of the moderated file. Valid values:
object: The file is in COS.
objecturl: The file is specified by URL, which can be an external URL.
API: The file is moderated through API call.
Moderation status
Success: The moderation succeeded.
Failed: The moderation failed.
Error code
If the moderation status is Failed, the error cause will be returned through an error code. For more information, see Error Codes.
Error message
Error message of the moderation failure.
bizType
Unique ID of the moderation policy, by which you can find the currently configured moderation policy.

Manually moderating results

After you filter results, the Moderation Details page will show filtered results, and you can perform the following operations on the filtered results:
Block an image or set its status to Normal.
Click a moderated image to view its details.

Automatic Moderation

Setting Live Stream Moderation

Last updated：2024-01-06 15:31:10

Overview

This document describes how to use the live stream moderation feature via the Cloud Object Storage (COS) console. Moderation scenarios for live stream moderation include pornographic, illegal, and advertising content inspection. After you enable the live stream moderation feature, newly generated video streams from specific live domains will be automatically moderated. You can view real-time moderation results through the console or callback information.

Prerequisites

Supported live stream duration: less than 5 hours.
Supported live stream resolutions: Up to 1920×1080 (1080p).
Supported live stream protocols: Mainstream protocols such as RTMP, HLS, HTTP, and HTTPS.

Operations

1. Log in to the COS console. On the Bucket List page, select the required bucket to open the bucket management page.
2. From the left sidebar, select Sensitive Content Moderation and then click Auto Audit Configuration > Live streaming moderation.
3. Click Add Live streaming moderation configuration to open the live stream moderation configuration page, and set the following configuration items:
Select a moderation policy: Choose the moderation policy you have configured. Different moderation policies correspond to different categories. You can customize the policy to personalize moderation based on scenes. Pornographic, illegal and advertising content moderation is supported, and you can select one or more detection scenes. You can go to Setting Moderation Policy to see how to configure a moderation policy.
Risk library associated: Risk libraries associated with the moderation policy.
Moderation scene: You can filter scenes configured in the moderation policy, such as pornography or advertising. You can select one or multiple scenes.
Play Domain: The playback domain name used in your live stream services. We will pull out the video stream from this domain for moderation.
Authentication key: If you have configured playback authentication in your live stream services, you need to enter the corresponding authentication key for your playback domain.
Callback configuration: If callback is activated, corresponding live stream moderation results will be delivered to you. You need to choose the moderation type and content for callback while configuring the callback URL. For further details, see Live Stream Moderation Callback Content.
Dedicated callback security configuration: To ensure the security of the CI callback information reception, we will generate a default moderation authentication token, which allows us to verify the source of the callback information. You can also customize and modify this token.
4. If you have an ongoing live stream services, you need to specify the corresponding configurations with existing callbacks, to ensure your existing stream initiation and stream termination callbacks are unaffected. If you use the live stream service for the first time, you can skip this step:
Existing Stream Initiation Callback: Specify your existing Live Stream Initiation Callback. We will forward the callback information to you.
Existing Stream Termination Callback: Specify your existing Live Stream Termination Callback. We will forward the callback information to you.
5. Click Save. The system will generate a callback address dedicated for live stream moderation. Specify this dedicated callback address for both your live stream initiation and termination callback in your CSS product. We will trigger the moderation automatically according to the stream initiation callback, and cease the moderation in line with the stream termination callback.
Click to go to the CSS console
6. After the preceding operations are completed, initiate a live stream to automatically commence the moderation process.


Setting Image Moderation

Last updated：2024-01-06 15:31:10

Overview

This document describes how to use the image moderation feature in the console. After you enable image moderation, new images uploaded to a bucket will be automatically moderated, and the identified non-compliant content can be automatically blocked (by denying public read access to the content).
You can also scan historical image files in COS to detect pornographic, illegal, and advertising content. For more information, see Single Image Moderation.
Note:
Supported image formats: PNG, JPEG, JPG, BMP, WEBP, GIF.
Supported image file size: < 32 MB. To moderate images larger than 5 MB in size, you need to enable the large image moderation feature.
Supported image dimensions: > 20 * 20 px (width * height).

Flowchart





Directions

1. Log in to the COS console. On the Bucket List page, click the target bucket to enter the bucket details page.
2. On the left sidebar, select Sensitive Content Moderation > Automatic Moderation Configuration and click Image Moderation.
3. Click Add Automatic Image Moderation Configuration and set the following configuration items:
Moderation Scope: You can select The whole bucket or Specified Range.
Path: If you select Specified Range, enter the path of the images to be moderated.
Example 1: To moderate files in the test directory, set this field to test/.
Example 2: To moderate files prefixed with 123, set this field to 123.
Note:
You can add multiple moderation configurations, but the paths cannot be duplicate or have inclusion relationships. If you have configured to moderate the entire bucket, you cannot add a moderation configuration for a specific path in the bucket.
Moderation Suffix: The following image formats can be moderated: JPG, JPEG, PNG, BMP, WEBP, and GIF. Two options are supported: Smart and Without suffix.
Note:
Smart: After you select this option, the system will intelligently determine whether a file is an image based on its file extension and content.
Large Image Moderation: The image moderation feature can moderate images below 5 MB in size. For larger images, you can enable the large image moderation feature. Then, the backend will compress images (up to 32 MB) first before moderating them.
Moderation Policy: Select a moderation policy. You can create different policies for refined moderation. Moderation scene options include Pornographic, Illegal, and Advertisement, and you can select one or multiple options. For detailed directions on how to configure a moderation policy, see Setting Moderation Policy.
Moderation Scene: It displays the default scene or the scene that you configure in the moderation policy. You can select the target scene as needed.
File block configuration: You can enable this service to authorize CI to perform automatic or human moderation and block the identified non-compliant files by denying public read access to them.
Block mode: The following two block modes are supported:
Change the file ACL to private read: Doing so actually blocks the file. Then, a 403 status code will be returned when the file is accessed again, indicating that access is denied. For more information on file permissions, see ACL.
Transfer the file to the backup directory: Doing so actually blocks the file. Then, a 404 status code will be returned when the file is accessed again, indicating that the file does not exist. The backup directory is automatically generated by the backend at audit_freeze_backup/increment_audit in the current bucket.
Block Type: You can select a block type and mechanism. Machine moderation and block is selected by default. If you select Human moderation and block, Tencent Cloud security team will review suspiciously sensitive images identified during machine moderation. You can select the image score range for blocking (by specifying an integer between 60 and 100; the greater the score, the more sensitive the image).
Callback: After callback is enabled, you will receive image moderation results. You need to select the moderation type, callback content, callback URL, and image domain name. If you select Custom Callback Threshold, you need to set the score range of the images for callback. After the callback URL is set, CI will send the default callback message to the set URL to check whether it can receive callback messages normally. For more information on callback, see Image Moderation Callback Content.
4. After completing the configuration, click Save. Images uploaded subsequently will be moderated.

Notes

1. Image moderation adopts a scoring mechanism, with a score between 0 and 100 returned for each image.
2. The confirmed part refers to the images that are confirmed to be normal or sensitive (with a score in the range of [0,60] or (90,100]). For such images, no manual intervention is required.
3. The uncertain part refers to the images that are suspected to be sensitive (with a score in the range of (60,90]). For such images, human moderation is recommended.

Setting Video Moderation

Last updated：2024-01-06 15:31:10

Overview

This document describes how to use the video moderation feature in the COS console. Detection scenarios include pornographic, illegal, and advertising information detection. After you enable video moderation, new videos uploaded to a bucket will be automatically moderated, and the identified non-compliant content can be automatically blocked (by denying public read access to the content).
You can also moderate historical videos stored in COS. For more information, see Submitting Video Moderation Job.
Note:
Video moderation is conducted by capturing video frames and checking screenshots. You can select the frame capturing method when you enable video moderation. For more information, see the directions below.
The following video formats can be moderated: MP4, AVI, MKV, WMV, RMVB, FLV, M3U8, MOV, M4V, and 3GP. The video size cannot exceed 5 GB, and the number of captured frames cannot exceed 10,000.

Flowchart





Directions

1. Log in to the COS console. On the Bucket List page, click the target bucket to enter the bucket details page.
2. On the left sidebar, select Sensitive Content Moderation > Automatic Moderation Configuration and click Video Moderation.
3. Click Add Automatic Video Moderation Configuration and set the following configuration items:
Moderation Scope: You can select The whole bucket or Specified Range.
Path: If you select Specified Range, enter the path of the videos to be moderated.
Example 1: To moderate files in the test directory, set this field to test/.
Example 2: To moderate files prefixed with 123, set this field to 123.
Note:
You can add multiple moderation configurations, but the paths cannot be duplicate or have inclusion relationships. If you have configured to moderate the entire bucket, you cannot add a moderation configuration for a specific path in the bucket.
Moderation Suffix: Supported video formats are MP4 (including other formats under the same MP4 format source: MPG, MPEG, MPE, DAT, VOB, 3GP), WMV (including other formats under the same WMV format source: ASF), RMVB (including other formats under the same RMVB format source: RM), and FLV (including other formats under the same FLV format source: F4V). If you select the Without suffix option, the system will check the Content-Type header of files without extensions to determine whether they are videos. Multiple options can be selected.
Moderation Policy: Select a moderation policy. You can create different policies for refined moderation. Moderation scene options include Pornographic, Illegal, and Advertisement, and you can select one or multiple options. For detailed directions on how to configure a moderation policy, see Setting Moderation Policy.
Moderate: Video image and video sound can be moderated. To moderate video sound, you need to select it in the moderation policy.
Note:
The Moderate option should be used together with the moderation policy. If video sound is not set in the policy, you cannot select Video sound here.
Moderation Scene: Moderation scene options include moderation for pornographic, illegal, and advertising content, and you can select one or multiple options.
Frame Capturing Rule: Video moderation is conducted by capturing video frames and checking screenshots. You can select Fixed Time, Fixed Frame Rate, or Fixed Quantity for video frame capturing.
Fixed Time: This option indicates to capture images at fixed intervals to moderate. You can set the time interval and the maximum number of frames per video.
Fixed Frame Rate: This option indicates to capture a fixed number of frames per second to moderate. You can set the number of frames captured per second and the maximum number of frames per video.
Fixed Quantity: This option indicates to moderate a fixed number of captured images for the entire video according to the average percentage.
File block configuration: You can enable this service to authorize CI to perform automatic or human moderation and block the identified non-compliant files by denying public read access to them. After enabling this service, you need to select the block type.
Block mode: The following two block modes are supported:
Change the file ACL to private read: Doing so actually blocks the file. Then, a 403 status code will be returned when the file is accessed again, indicating that access is denied. For more information on file permissions, see ACL.
Transfer the file to the backup directory: Doing so actually blocks the file. Then, a 404 status code will be returned when the file is accessed again, indicating that the file does not exist. The backup directory is automatically generated by the backend at audit_freeze_backup/increment_audit in the current bucket.
Block Type: You can select a block type and mechanism. Machine moderation and block is selected by default. If you select Human moderation and block, Tencent Cloud security team will review suspiciously sensitive videos identified during machine moderation.
Callback: After callback is enabled, you will receive video moderation results. You need to select the moderation type and callback content and set the callback URL. For more information, see Video Moderation Callback Content.
4. After completing the configuration, click Save. Videos uploaded subsequently will be moderated.

Notes

1. Video moderation adopts a scoring mechanism, with a score between 0 and 100 returned for each captured video frame.
2. The confirmed part refers to the videos that are confirmed to be normal or sensitive (with a score in the range of [0,60] or (90,100]). For such images, no manual intervention is required.
3. The uncertain part refers to the videos that are suspected to be sensitive (with a score in the range of (60,90]). For such images, human moderation is recommended.

Setting Audio Moderation

Last updated：2025-11-19 20:07:36

Overview

This document describes how to use the audio moderation feature in the console to check audio content for pornographic, illegal, and advertising information.
After you configure automatic audio moderation, new audios uploaded to a bucket will be automatically moderated, and the identified non-compliant content can be automatically blocked (by denying public read access to the content).
You can also moderate existing audios stored in COS. For more information, see Audio Moderation.
Note:
Audio moderation is billed by CI.
Supported audio formats: MP3, WAV, AAC, FLAC, AMR, 3GP, M4A, WMA, OGG, APE.
Supported audio bitrate: 128–256 Kbps.
Supported audio size: < 600 MB.
Maximum duration: 3 hours.
Audio moderation can recognize Mandarin and English.

Flowchart





Directions

1. Log in to the COS console. On the Bucket List page, click the target bucket to enter the bucket details page.
2. On the left sidebar, select Sensitive Content Moderation > Automatic Moderation Configuration and click Audio Moderation.
3. Click Add Automatic Audio Moderation Configuration and set the following configuration items:
Moderation Scope: Select the scope of audio files to be moderated, which can be the entire bucket, a specific directory, or a specific file prefix.
Moderation Suffix: The following audio formats can be moderated: MP3, WAV, AAC, FLAC, AMR, 3GP, M4A, WMA, OGG, and APE. You can select multiple formats.
Moderation Policy: Select a moderation policy. You can create different policies for refined moderation. If no policies have been configured, the default policy will be used. Moderation scene options include Pornographic, Illegal, and Advertisement, and you can select one or multiple options. For detailed directions on how to configure a moderation policy, see Setting Moderation Policy.
Moderation Scene: It displays the scene that you configure in the moderation policy. You can select the target scene as needed.
File block configuration: You can enable this service to authorize CI to perform automatic or human moderation and block the identified non-compliant files by denying public read access to them. After enabling this service, you need to select the block type and score range of audios to be blocked.
Block mode: The following two block modes are supported:
Change the file ACL to private read: Doing so actually blocks the file. Then, a 403 status code will be returned when the file is accessed again, indicating that access is denied. For more information on file permissions, see ACL.
Transfer the file to the backup directory: Doing so actually blocks the file. Then, a 404 status code will be returned when the file is accessed again, indicating that the file does not exist. The backup directory is automatically generated by the backend at audit_freeze_backup/increment_audit in the current bucket.
Block Type: You can select a block type and mechanism. Machine moderation and block is selected by default. If you select Human moderation and block, Tencent Cloud security team will review suspiciously sensitive audios identified during machine moderation. You can select the audio score range for blocking (by specifying an integer between 60 and 100; the greater the score, the more sensitive the audio).
Callback: After callback is enabled, you will receive moderation results. You need to select the moderation type and callback content and set the callback URL. For more information, see Callback Content.
4. After completing the configuration, click Save. Audios uploaded subsequently will be moderated.

Callback Content

If you enable callback, after audio moderation is completed, the system will send the following callback message to the callback URL:
{
    "code":0,
    "message":"success",
    "data":{
        "url":"",
        "result":1,
        "forbidden_status":1,
        "trace_id":"",
        "porn_info":{
            "hit_flag":1,
            "score":91,
            "label":""
        },
        "ads_info":{
            "hit_flag":0,
            "score":0,
            "label":""
        }
    }
}
Parameter
Description
Type
Required
forbidden_status
Block status. Valid values: 0 (normal); 1 (blocked).
Int
Yes
porn_info
Porn detection information, including moderation result, score, and detailed tags.
json
Yes
ads_info
Ad detection information, including moderation result, score, and detailed tags.
json
Yes
result
Recognition result for reference. Valid values: 0 (normal), 1 (sensitive), 2 (suspiciously sensitive).
Int
Yes
trace_id
jobid of the submitted moderation job.
String
Yes
url
The URL of the uploaded resource, including the domain name.
String
Yes
The moderation information (porn_info and ads_info) has the following sub-nodes:
Parameter
Description
Type
Required
hit_flag
Whether the moderation type is hit. Valid values: 0 (miss), 1 (hit), and 2 (suspected).
Int
Yes
label
Recognized audio tag.
String
Yes
score
Moderation score. Valid value ranges: 0–60 (normal); 60–90 (suspiciously sensitive); 90–100 (sensitive).
Int
Yes

Notes

1. Audio moderation adopts a scoring mechanism, with a score between 0 and 100 returned for each audio.
2. The confirmed part refers to the audios that are confirmed to be normal or sensitive (with a score in the range of [0,60] or (90,100]). For such audios, no manual intervention is required.
3. The uncertain part refers to the audios that are suspected to be sensitive (with a score in the range of (60,90]). For such audios, human moderation is recommended.

Setting Text Moderation

Last updated：2025-11-19 20:07:36

Overview

This document describes how to use the text moderation feature in the COS console. The feature can check text content for pornographic, illegal, advertising, and abusive information.
After you enable text moderation, new text files uploaded to a bucket will be automatically moderated, and the identified non-compliant content can be automatically blocked (by denying public read access to the content).
Note:
Text moderation is billed by CI.
Text moderation is billed by moderation times. Every 10,000 UTF-8 characters is counted as one moderation operation, and less than 10,000 characters are counted as 10,000 characters.
Currently, the text moderation feature supports TXT files and files without extensions, and the file size cannot exceed 1 MB.
Text moderation can recognize Mandarin and English.

Flowchart





Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the target bucket to go to the configuration page.
4. On the left sidebar, select Sensitive Content Moderation > Automatic Moderation Configuration and click Text Moderation.
5. Click Add Automatic Text Moderation Configuration and set the following configuration items:
Moderation Scope: You can select The whole bucket or Specified Range.
Path: If you select Specified Range, enter the path of the text files to be moderated.
Example 1: To moderate files in the test directory, set this field to test/.
Example 2: To moderate files prefixed with 123, set this field to 123.
Note:
You can add multiple moderation configurations, but the paths cannot be duplicate or have inclusion relationships. If you have configured to moderate the entire bucket, you cannot add a moderation configuration for a specific path in the bucket.
Moderation Suffix: Options include TXT, HTML, and Without suffix.
Moderation Policy: Select a moderation policy. You can create different policies for refined moderation. Moderation scene options include Pornographic, Illegal, and Advertisement, and you can select one or multiple options. For detailed directions on how to configure a moderation policy, see Setting Moderation Policy.
Moderation Scene: Moderation scene options include moderation for pornographic, advertising, and abusive content, and you can select one or multiple options.
File block configuration: You can enable this service to authorize CI to perform automatic or human moderation and block the identified non-compliant files by denying public read access to them.
Block mode: The following two block modes are supported:
Change the file ACL to private read: Doing so actually blocks the file. Then, a 403 status code will be returned when the file is accessed again, indicating that access is denied. For more information on file permissions, see ACL.
Transfer the file to the backup directory: Doing so actually blocks the file. Then, a 404 status code will be returned when the file is accessed again, indicating that the file does not exist. The backup directory is automatically generated by the backend at audit_freeze_backup/increment_audit in the current bucket.
Block Type: You can select a block type and mechanism. Machine moderation and block is selected by default. If you select Human moderation and block, Tencent Cloud security team will review suspiciously sensitive text files identified during machine moderation.
Callback: After callback is enabled, you will receive moderation results. You need to select the moderation type and callback content and set the callback URL. For more information, see Callback Content.
6. After completing the configuration, click Save. Text files uploaded subsequently will be moderated.

Callback Content

After callback is enabled, the system will send the following default callback message to the set URL to check whether it can receive callback messages normally:
{
   "code": 0,
   "data": {
       "forbidden_status": 0,
       "porn_info": {
           "hit_flag": 0,
           "label": "",
           "count": 9
       },
       "result": 0,
       "trace_id": "test_trace_id",
       "url": "test_text"
   },
   "message": "Test request when setting callback url"
}
Note:
If the callback option is selected, text files frozen by Tencent Cloud will be returned to you, with public read access to them denied.
The callback URL must begin with "http" or "https" and return a 200 status code by default before it can be used. Check it before saving the settings.
The callback URL will take effect in 30 minutes.
After the callback URL takes effect, when an uploaded text file hits moderation rules, the system will call back the URL by default and send a standard HTTP POST notification message to it. The HTTP packet is as follows:
Parameter
Description
Type
Required
forbidden_status
Block status. Valid values: 0 (normal); 1 (blocked).
Int
Yes
porn_info
Porn detection information, including moderation result, score, and detailed tags.
json
Yes
ads_info
Ad detection information, including moderation result, score, and detailed tags.
json
Yes
result
Recognition result for reference. Valid values: 0 (normal), 1 (sensitive), 2 (suspiciously sensitive).
Int
Yes
trace_id
jobid of the submitted moderation job.
String
Yes
url
The URL of the uploaded resource, including the domain name.
String
Yes
illegal_info
Illegal information detection information, including moderation result, score, and detailed tags.
json
No
abuse_info
Abuse detection information, including moderation result, score, and detailed tags.
json
No
The moderation information (porn_info, ads_info, illegal_info, and abuse_info) has the following sub-nodes:
Parameter
Description
Type
Required
hit_flag
Whether the moderation type is hit.
Int
Yes
label
Recognized text tag.
String
Yes
count
Text file callback parameter, which is the number of sensitive text segments that hit the moderation scene.
Int
Yes
Below is a sample callback:
{	
    "code":0,
       "message":"success",
       "data":{
          "url":"xxxxxxxxxxxxxxx",
          "result":1,
          "forbidden_status":1,
          "trace_id":"xxxxxxxxxxxxxxx",
          "porn_info":{
              "hit_flag":1,
              "label":"Obscene",
              "count":3
           },
    },
}


Setting Document Moderation

Last updated：2024-01-06 15:31:10

Overview

This document describes how to use the document moderation feature in the console to check file content for pornographic, illegal, and advertising information.
After you configure automatic document moderation, new documents uploaded to a bucket will be automatically moderated, and the identified non-compliant content can be automatically blocked (by denying public read access to the content).
You can also moderate existing documents stored in COS. For more information, see Document Moderation.
Note:
The document moderation feature leverages the document conversion capability to convert each page of a document into an image for moderation.
Document moderation is billed by CI.
Currently, document types supported for moderation include:
Presentation files: PPTX, PPT, POT, POTX, PPS, PPSX, DPS, DPT, PPTM, POTM, PPSM.
Text files: DOC, DOT, WPS, WPT, DOCX, DOTX, DOCM, DOTM.
Spreadsheet files: XLS, XLT, ET, ETT, XLSX, XLTX, CSV, XLSB, XLSM, XLTM, ETS.
PDF.
Other files: TXT, LOG, HTM, HTML, LRC, C, CPP, H, ASM, S, JAVA, ASP, BAT, BAS, PRG, CMD, RTF, XML.
A spreadsheet file may be split into multiple pages, with multiple images generated.
The input file size cannot exceed 200 MB.
The number of pages in the input file cannot exceed 5,000.

Flowchart





Directions

1. Log in to the COS console.
2. On the Bucket List page, click the target bucket to enter the bucket details page.
3. On the left sidebar, select Sensitive Content Moderation > Automatic Moderation Configuration and click Document Moderation.
4. Click Add Automatic Document Moderation Configuration and set the following configuration items:
Moderation Scope: Select the scope of documents to be moderated, which can be the entire bucket, a specific directory, or a specific file prefix.
Moderation Suffix: Select one or multiple options for Document Format, including presentation, text, spreadsheet, and PDF.
Moderation Policy: Select a moderation policy. You can create different policies for refined moderation. If no policies have been configured, the default policy will be used. Moderation scene options include Pornographic, Illegal, and Advertisement, and you can select one or multiple options. For more information on how to configure a moderation policy, see Setting Moderation Policy.
Moderation Scene: It displays the scene that you configure in the moderation policy. You can select the target scene as needed.
File block configuration: You can enable this service to authorize CI to perform automatic or human moderation and block the identified non-compliant files by denying public read access to them. After enabling this service, you need to select the block type and score range of files to be blocked.
Block mode: The following two block modes are supported:
Change the file ACL to private read: Doing so actually blocks the file. Then, a 403 status code will be returned when the file is accessed again, indicating that access is denied. For more information on file permissions, see ACL.
Transfer the file to the backup directory: Doing so actually blocks the file. Then, a 404 status code will be returned when the file is accessed again, indicating that the file does not exist. The backup directory is automatically generated by the backend at audit_freeze_backup/increment_audit in the current bucket.
Block Type: You can select a block type and mechanism. Machine moderation and block is selected by default. If you select Human moderation and block, Tencent Cloud security team will review suspiciously sensitive files identified during machine moderation. You can select the file score range for blocking (by specifying an integer between 60 and 100; the greater the score, the more sensitive the file).
Callback: After callback is enabled, you will receive moderation results. You need to select the moderation type and callback content and set the callback URL. For more information, see Document Moderation Callback Content.
5. After completing the configuration, click Save. Documents uploaded subsequently will be moderated.

Notes

1. Document moderation adopts a scoring mechanism, with a score between 0 and 100 returned for each output image.
2. Depending on the score range, the moderation result can be a sensitive image, suspiciously sensitive image, or normal image.
The score range of sensitive images is ≥ 91.
The score range of suspiciously sensitive images is 61–90. Such images cannot be accurately identified as sensitive, so human moderation is recommended to ensure their content security.
The score range of normal images is ≤ 60. Such images are determined as normal by the system.

Historical data moderation

Configuring Historical Data Moderation Tasks

Last updated：2024-03-25 15:00:04

Overview

This document describes how to use the historical data moderation feature of Cloud Object Storage (COS) via the console. You can create a historical data moderation task to conduct a batch moderation on your images, videos, audios, texts, and files all at once.

Directions

Creating a Moderation Task

1. Log in to the COS console.
2. Select Bucket List in the left navigation bar to open the bucket management page.
3. Navigate to the target bucket and click on its name to open the bucket configuration page.
4. In the left navigation bar, choose Sensitive Content Moderation > Existing File Moderation to open the historical data moderation page.
5. Click Create Moderation Task.
6. In the Scan Range page, you can moderate your files as needed through a specified scan scope:



The scan scope include Bucket File List, COS Inventory Report, and URL List File:
Bucket File List: You can select files within the current bucket for moderation. The scan range can either be the file upload time or prefix.
COS inventory report: You can choose the inventory list generated by the COS Setting Inventory feature to scan, and put the inventory list file in the current bucket.
URL list file: You can choose a designated URL list file to scan. Currently, TXT files with one URL listed per line are supported.
7. Click Next.
8. On the Moderation Policy page, activate the review policy as needed, configure the corresponding file type for moderation and the review scenario type, and click Next.



Review Image:
Moderation Suffix: Image moderation supports the following formats: JPG, JPEG, PNG, BMP, WEBP, and GIF.
Note:
Intelligent suffix recognition can recognize some special suffixes in addition to the preceding six common suffixes.
Large Image Moderation: Image moderation only supports images up to 5 MB. For images exceeding the size limit, you can enable the large image moderation feature. Then the system compresses the images before the moderation process. This feature can compress images up to 32 MB.
Note:
The large image moderation feature incurs basic image processing fees. For detailed prices, see Basic Image Processing Costs.
Moderation Policy Selection: Please select the moderation policy you have configured (if no policy is configured, the system's default policy can be used). Each moderation policy corresponds to a category, and you can customize moderation scenes through policy customization. Options for moderation include pornographic content, illegal activity or non-compliant behavior, and advertising content, and you can select one or multiple scenes for moderation. For more information on how to configure a moderation policy, see Setting Moderation Policy.
Daily Moderation Limit: You can configure the maximum number of images that can be moderated per day, By default, there is no upper limit.
Moderation Scene: Either the default scenes or those you have configured in your moderation policy are displayed here. You can select the scene categories you wish to moderate.
Review Video:
Moderation Suffix: Video moderation supports the following formats: MP4, AVI, MKV, WMV, RMVB, FLV, M3U8, and so on.
Moderation Policy Selection: Please select the moderation policy you have configured. Each moderation policy corresponds to a category, and you can customize moderation scenes through policy customization. Options for moderation include pornographic content, illegal activity or non-compliant behavior, and advertising content, and you can select one or multiple scenes for moderation. For more information on how to configure a moderation policy, see Setting Moderation Policy.
Daily Moderation Limit: You can configure the maximum number of video files that can be moderated per day. By default, there is no upper limit.
Moderation Scene: Moderation scene options include moderation for pornographic content, illegal, and advertising content. You can select one or multiple moderation scenes.
Moderation Content: Both video images and audio can be moderated.
Frame Capturing Rule: Frame capturing is the basis for video moderation, which is implemented through the moderation of the image captured from the video. Fixed Time, Fixed Frame Rate, and Fixed Quantity are supported for frame capturing moderation.
Fixed Time: Images are captured at fixed intervals for moderation. You can set the time interval and the maximum number of frames captured per video.
Fixed Frame Rate: A fixed number of frames are captured per second for moderation. You can set the frame capture rate per second and the maximum number of frames captured per video.
Fixed Quantity: A predetermined quantity of images are captured from full-length videos, based on an average percentage. You can set the maximum number of frame captured per video.
Note:
The configuration of frame capturing rules will impact the outcome of the moderation.
Review Audio:
Moderation Suffix: Audio moderation supports the following formats: MP3, WAV, AAC, FLAC, AMR, 3GP, M4A, WMA, OGG, and APE.
Moderation Policy Selection: Please select the moderation policy you have configured (if no policy is configured, the system's default policy can be used). Each moderation policy corresponds to a category, and you can customize moderation scenes through policy customization. Options for moderation include pornographic content, illegal activity or non-compliant behavior, and advertising content, and you can select one or multiple scenes for moderation. For more information on how to configure a moderation policy, see Setting Moderation Policy.
Daily Moderation Limit: You can set the maximum quantity of audio files to be moderated per day. By default, there is no upper limit.
Moderation Scene: Moderation categories display the scenes that you configured in the moderation policy, and you can select the desired categories for moderation.
Review Text:
Moderation Suffix: Text files with the suffix TXT or no suffix can be moderated.
Moderation Policy Selection: Please select the moderation policy you have configured (if no policy is configured, the system's default policy can be used). Each moderation policy corresponds to a category, and you can customize moderation scenes through policy customization. Options for moderation include pornographic content, illegal activity or non-compliant behavior, and advertising content, and you can select one or multiple scenes for moderation. For more information on how to configure a moderation policy, see Setting Moderation Policy.
Daily Moderation Limit: You can set the maximum amount of text files moderated per day. By default, there is no upper limit.
Moderation Scene: Moderation categories display the scenes that you configured in the moderation policy, and you can select the desired categories for moderation.
Review More:
Moderation Suffix: The file formats supporting moderation include presentations, text files, spreadsheets, PDFs, among others. You can select multiple formats.
Moderation Policy Selection: Please select the moderation policy you have configured (if no policy is configured, the system's default policy can be used). Each moderation policy corresponds to a category, and you can customize moderation scenes through policy customization. Options for moderation include pornographic content, illegal activity or non-compliant behavior, and advertising content, and you can select one or multiple scenes for moderation. For more information on how to configure a moderation policy, see Setting Moderation Policy.
Daily Moderation Limit: You can configure the maximum number of files that can be moderated per day. By default, there is no upper limit.
Moderation Scene: The scenes you have configured in your selected moderation policy are displayed here. You can select the scenes you want to moderate.
9. In the Block Policy interface, configure the block policy, and then click Next. Enabling the block policy authorizes Cloud Infinite (CI) services to conduct automatic moderation or manual re-moderation on certain types of files, forbidding public read access to any identified non-compliant content.



Image blocking and human review settings: You can set values (that is, integers from 60 to 100) based on different moderation types. The images will be blocked when the set value range is reached. By default, direct blocking is selected. If you select manual re-moderation, a secondary review on the image will be carried out by a professional security team.
Note:
The moderation results are classified into confirmed sensitive, likely sensitive, and normal categories based on the moderation score.
The score range for confirmed sensitive images is ≥ 91.
The score range for potentially sensitive images is from 61 to 90. The system may not accurately identify these images as sensitive, thus it's recommended that users should conduct manual re-moderation to ensure the anti-spam nature of the images.
The score range for normal images is ≤ 60. The system considers these images as normal.
CDN Cache Purge after Blocking: When it is enabled, the corresponding domain's CDN cache data will be purged while the COS origin file is blocked.
Block Mode: The following two blocking modes are supported.
Change the file ACL to private: Files are blocked through the modification of the file's access permission to private read. When this mode is used, a re-attempt to access the file would return a 403 status code, indicating that access to the file is not allowed. For information about file permissions, see ACL.
Transfer the file to the backup directory: Files are block through the move of the file to the backup directory. When this mode is used, a re-attempt to access the file returns a 404 status code, indicating that the file does not exist. The backup directory is automatically generated by the background, located in the path of the current bucket: audit_freeze_backup/increment_audit.
10. On the Moderation Results page, set up the callback for moderation results and click Next.
After enabling callback settings, we will direct the moderation results to your designated callback address. You need to set the callback type, callback content, and callback URL.



Callback Scenario: Based on the moderation policy you have set, options include pornographic content, illegal and non-compliant content, advertisement moderation, illegal content, and verbal abuses.
Callback Mode: You have chosen either Lite or Detailed mode.
Callback Content: Options include callback only for non-compliant files, callback only for blocked files, and callback for all files. Re-moderation is also supported for files that failed the first moderation.
Callback URL: The callback URL must by default return a 200 status code.
Callback URL Protocol: You can choose either enforced HTTP or HTTPS.
11. After checking all task configurations, click Create to complete task creation.

Viewing Task Results

On the Existing File Moderation page, you can perform different operations based on the task status.


When the task status is under review, you can click Task Configuration or Terminate Task.
When the task status is Job completed, you can view Moderation Details or Result Statistics.
Moderation Details: Only moderation details from the past one month can be viewed. After clicking this option, you will be redirected to the moderation page where you can export moderation results, conduct manual moderation, and so on. For specific operation instructions, see Moderation Details.
Result Statistics: This page displays the statistical results of the moderation task. If you have any doubts about the moderation results, you can go to the Moderation Details page on the control panel to view the specific moderation contents.


Setting Moderation Policy

Last updated：2024-01-06 15:31:10

Overview

When you use the content moderation service, you can specify a moderation category through a moderation policy to moderate content in custom scenes. For different file types, COS provides different scenes for your choice, making it easier for you to customize moderation policies that suit your business.
Currently, the file types and corresponding moderation scenes supported by COS are as follows:
Supported File Types
Moderation Scenes
Specific Moderation Categories
Image moderation policy, video moderation policy, audio moderation policy, text moderation policy, file moderation policy,Webpage Moderation policy
Pornographic content
Sexually suggestive/vulgar behaviors
Genital nudity/Sexual behaviors
Sex toys
Sexy content
Pornographic text moderation
Advertising content
QR code and barcode recognition
Logo detection and recognition
Advertising text moderation
Other non-compliant content
Fire, explosion, bloody scene, etc.

Note:
WEBP images involving advertising QR codes cannot be moderated currently.

Directions

Default policy

Each moderation type has a default moderation policy, which is configured by Tencent Cloud based on your historical moderation conditions. If you have never used the moderation service, the policy developed by algorithm experts based on models for multiple industries will be used by default, which will be suitable for most content security requirements.
Note:
The default policy can be viewed and edited but not deleted.

Custom policy

If the default policy cannot meet your business needs, or you have multiple scenes that require different moderation policies, you can create custom policies as needed.
1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the target bucket to enter the configuration page.
4. Select Content Moderation > Moderation Policy on the left sidebar.
5. Create a moderation policy based on your business needs. Currently, you can create policies for image moderation, video moderation, audio moderation, text moderation, and file moderation. Take the image moderation policy as an example: i. Click Create Image Moderation Policy and enter the policy name. ii. Select the category you need to moderate under the moderation type, where OCR pornographic text moderation indicates to moderate image content based on OCR.
iii. Click Save to create the policy.
6. After the moderation policy is created, the backend will automatically generate a unique Biztype value.
7. You can view or edit the created moderation policy; however, you cannot change its name and Biztype value.

Using moderation policy

After the moderation policy is created, configure automatic moderation, create a historical data moderation job, or call a content moderation API. You can select the corresponding policy to moderate content in custom categories.

Automatic moderation

Create an automatic moderation configuration in the console, and you can manually select a moderation policy.

Historical data moderation

Create a historical data moderation job in the console, and you can manually select a moderation policy.

Moderation APIs

When calling an API for content moderation, you can manually pass in the Biztype value to the API, which you can view in the list of moderation policies in the console. If you don't pass it in, the default policy will be used.

Setting the Custom Image Risk Library

Last updated：2024-03-25 15:01:26

Overview

The custom image risk library is designed to help you manage images or keywords that need to be moderated. With risk library customization, you can configure the images or keywords to allow or block them. This can be applied to all moderation scenes.

Image Risk Library

You can use an image risk library to manage images that need to be blocked or allowed, to address unexpected control requirements.
Six preset image risk libraries are available in the system:
Normal image library: The returned moderation result will be normal for images that match those in this library.
Pornographic and illegal image library: The returned moderation result will be the corresponding violation tag for images that match those in the library.
Creating new image risk library is not supported. You can add sample images to the existing default libraries.
Note:
The image risk library is effective by account. After you add sample images to the gallery in any bucket under the same account, these images will automatically take effect in all your buckets and all your moderation policies.
A single image library can contain a maximum of 10,000 sample images.
Some specific images may fail to be added to the image library. In this case, please contact us for assistance.

Text Risk Library

You can use the text risk library to manage text to be blocked or allowed, in order to address sudden control requirements.
The text risk library contains the following types:
Pre-defined text library: It is a text library with pre-defined policies for you, which includes:
Normal text library: If a keyword in the library is matched, the moderation result will be returned as normal.
Pornographic text library, illegal text library, and other text libraries: If a keyword in the library is matched, the moderation result will be returned as the corresponding violation tag.
Note:
The pre-defined text library takes effect by account. After you add sample keywords to any pre-defined text library for the same account, these keywords will automatically take effect in all your storage buckets and all moderation policies.
Up to 10,000 sample keywords can be added to a pre-defined text library.
Custom text library: It is the text library you create, where you can add samples of various types of violations. If the text under moderation matches a keyword in the library, it will be marked with the corresponding tag based on the defined library policy.
Note:
The custom text library needs to be associated with a moderation policy and only takes effect within the associated moderation policy.
Up to 2,000 sample keywords can be added to a single custom text library.

Directions

1. Log in to the Cloud Object Storage console. On the Bucket List page, select the required bucket to open the bucket management page.
2. On the left navigation bar, choose Sensitive Content Moderation > Custom Risk Library.
3. On the Custom Risk Library page, three tabs are displayed: Image Risk Library, Text Risk Library, and Business Field Risk Library. For instructions on operating the Business Field Risk Library, see Setting the Business Field Risk Library.

4. The operations on Image Risk Library and Text Risk Library are as follows:
Image Risk Library
To handle incorrect moderation, you can add image clean samples, to ensure the image moderation results return as normal:
Find the normal image library (pre-defined) in the list and click Manage to the right of the library to access the Image Risk Library page.
On the Image Risk Library page, you can perform the following operations:

Viewing the image library policy: The policy for the normal image library is normal.
Viewing samples: Check the number of samples added to the image library.
Adding samples: You can add selected images to the library as samples.
Deleting samples: You can delete sample images from your image library.
For scenarios of missed moderation, add image block samples so that the image moderation results return as sensitive:
In the list, find the image library for with the returned sensitive type is desired. If you add an image to the pornographic image library (pre-defined), the moderation result will be tagged as pornography. Click Manage on the right of the library to open the Image Risk Library page.
On the Image Risk Library page, you can perform the following operations:

Viewing the image library policy: The policy for the pornographic image library is sensitive.
Viewing samples: Check the number of samples added to the image library.
Adding samples: You can add specific images to the library as samples.
Deleting samples: You can delete sample images from your image library.
Text Risk Library
You can directly add keywords into the pre-defined text library or create a custom text library to add keywords:
Keywords added to the pre-defined text library will take effect for all moderation policies:

If you want to add clean sample keywords, find the normal text library (pre-defined) in the list, and then click Manage on the right side of the text library to open the Text Risk Library page.
On the Text Risk Library page, you can perform the following operations:

Viewing the text library policy: The policy of the normal text library is normal.
Viewing samples: Check the number of samples added to the text library.
Adding samples: You can add specific keywords into the text library as samples.
Deleting samples: You can delete keywords from the text library.
Creating a custom text library: The custom text library needs to be associated with a moderation policy. In moderation operations, only the text library associated with moderation policies takes effect:
Click Create Custom Text Library, and in the window, specify the Text Library Name, and select the Text Library Policy and Match Type:
Text Library Policy: When a keyword sample in the text library is matched, you can choose sensitive or suspected as the returned moderation result.
Match Type: You can choose an exact match or a fuzzy match. For a fuzzy match, variants of the entered keyword can be detected to match similar words. Split words, homographs, homophones, simplified or traditional forms, case differences, and numeral words are supported.
After the Custom Text Library has been created, find the newly created library in the list, and click Manage on the right side of the selected library to open the Text Risk Library page.
On the Text Risk Library page, you can perform the following operations:

Viewing the text library policy: The custom text library policy can be sensitive or suspected.
Viewing samples: Check the number of samples added to the text library.
Adding samples: You can add specific keywords into the text library as samples.
Deleting samples: You can delete keywords from the text library.
5. After the risk library is configured, if samples from the risk library are encountered while you use the content moderation feature, they will be automatically allowed or blocked according to the risk policy.


Setting the Business Field Risk Library

Last updated：2024-03-25 15:01:26

Overview

Content moderation supports adding the business field UserInfo to tag the data for moderation, including nickname information, account details, room number, and so on. For more information, see API Documentation.
This allows you to manage users or IP addresses that need to be blocked or allowed. For example, you can blacklist a certain nickname so that data entries containing that nickname field will be blocked by default.

Differences between the Business Field Risk Library and the Image and Text Risk Library

The business field risk library manages the business fields that you add when you call the moderation API, while the image and text risk library manages the content of images or text.
The business fields that can be added to the risk library include:
Field Name
Description
TokenId
It usually indicates account information, limited to 128 bytes in length.
Note
Up to 10,000 fields can be added to a business field risk library.

Operations

1. Log in to the Cloud Object Storage console. On the Bucket List page, select the required bucket to open to the Bucket List page.
2. On the left navigation bar, select Sensitive Content Moderation > Custom Risk Library.
3. On the Custom Risk Library page, select the Business Field Risk Library tab, and then click create business field risk library.


4. Enter the related content:
Library name: Up to 32 Chinese and English characters, digits, and underscores are supported.
Match field: Select the business field to be controlled. Currently, only TokenId is supported.
Library type: Select the corresponding library type. When the content in the library is matched, the returned moderation result will be the corresponding type. The blocklist result can be sensitive or suspected, and the allowlist result is normal.
Moderation policy associated: A business field risk library requires an associated policy and is effective only for the associated policy.
5. Click OK to completing creating the business field risk library.
6. After the library is created, you can manage, edit, or delete the library.
7. Click Manage to open the management interface of the current library. You can add or delete content on this page. The content here corresponds to the business field. For example, if you choose TokenId as the matching field (TokenId usually indicates the account ID), then the content here is the specific account ID, such as 12345678 and 1008787.

8. Click Edit to modify the account settings: The library name and associated moderation policy can be modified, but the matching field and library type cannot be modified.


Smart Toolbox

Last updated：2024-01-06 14:48:14

Overview

The smart toolbox provides tools for processing almost all types of multimedia files, such as image watermark, image compression, and audio/video/file format conversion. It presents various capabilities of COS in the form of convenient and quick tools in the console for you to use with speed and easy.

Notes

All the capabilities in the smart toolbox are implemented based on the data processing APIs of COS. Using them is equivalent to calling such APIs. You need to keep the following in mind during use:
When you use the smart toolbox for the first time, the CI service will be activated for you free of charge.
Whenever a processing operation is completed through the toolbox, it is equivalent to calling a data processing API once and incurs processing fees as detailed below:
Tool
Fees
Image compression (WebP or JPEG format), image editing, image watermark, and image information
Basic image processing
Image compression (AVIF or HEIF format)
Image advanced compression
Audio/Video processing
Media processing
AI processing
Content recognition
Online file preview
File preview

Tools

The smart toolbox currently contains the following tools:
Category
Tool
Image processing
Audio/Video processing
AI processing
File processing


Image processing

1. Image compression

The image compression tool implements the process of downsizing an image as much as possible without sacrificing quality so that it can be stored at a lower cost and accessed more quickly.

Directions

1. Log in to the COS console and click Smart Toolbox on the left sidebar.
2. On the Smart Toolbox page, select Image Processing > Image Compression.
3. In the image upload area, select an existing image in the bucket and add it to the tool.
4. After the image is added, the tool will automatically process it, and you can see the size of the compressed image on the left. You can also download the compressed image.

2. Image editing

The image editing tool offers capabilities of image cropping, rotation, scaling, sharpening, brightness adjustment, and contrast adjustment.

Directions

After adding an image as instructed in steps 1–3 in the Image compression section, use tools on the left to make corresponding adjustments.

3. Image watermark

The image watermark tool allows you to add an image or text to another image in the form of watermark.

Directions

Image watermark: After adding an image as instructed in steps 1–3 in the Image compression section, click Image Watermark on the left, select an existing image in the bucket, adjust the margins, and click Generate Watermark.
Text Watermark: Click Text Watermark on the left, enter the watermark text, adjust the margins, font, and font size, and click Generate Watermark.

4. Image information

The image information tool lists the format, size, and MD5 information of an image.

Directions

After adding an image, click Information on the left to view the information of the image.

Audio/Video processing

1. Audio/Video format conversion

The audio/video format conversion tool can convert your audio/video files into MP4, MP3, MOV, AVI, MKV, and other formats. It allows you to set different parameters such as video resolution and audio bitrate during conversion to adapt to different terminals and network environments.

Directions

1. Log in to the COS console and click Smart Toolbox on the left sidebar.
2. On the Smart Toolbox page, select Audio/Video Processing > Audio/Video Format Conversion.
3. Select an audio/video file and add it to the tool. You can select existing files in the bucket or local files, but if you select local files, you need to upload them to the bucket, as the tool can convert files in the bucket only.
4. After selecting an audio/video file, you need to select the conversion parameters as detailed below: The following is an example of converting a video with parameters of MP4 format, H.264 codec, 720 * proportional height resolution, and 1024 Kbps bitrate:
Video format: MP4
Video codec: H.264
Video resolution: 720 (width) * proportional height of the input video
Video bitrate: 1024 Kbps
5. After selecting parameters, select the name and location of the output video in the bucket and click Finish.
6. Click Start Transcoding and wait for the transcoding to complete.
7. After transcoding is completed, the input and output videos are displayed in the left and right of the video display area respectively for you to directly compare and check the transcoding effect. You can copy the output video link or directly download the output video by clicking buttons in the bottom-right corner.

2. Top speed codec transcoding

Top speed codec transcoding leverages deep learning algorithms to convert an input video to an output video with higher definition, lower noise, and higher frame rate by reducing the compression and texture distortion of the input video.

Directions

The steps of top speed codec transcoding are similar to those of audio/video file conversion, except that the output video transcoded by top speed codec is smaller and clearer.

3. Video frame capturing

With the video frame capturing tool, you can capture any frame in a video and save it as an image.

Directions

1. Add a video to the tool as instructed in steps 1–3 in the Audio/Video format conversion section and select frame capturing parameters.
2. After the video is added, the tool will automatically start capturing frames and display the captured frames on the right.

4. Video-to-animated image conversion

The video-to-animated image conversion tool can convert your video into a GIF or WebP animated image.

Directions

1. Add a video to the tool as instructed in steps 1–3 in the Audio/Video format conversion section and select video-to-animated image conversion parameters.
2. After the video is added, the tool will automatically start converting and display the output animated image on the right.

5. Intelligent thumbnail

By intelligently recognizing and analyzing the characteristics of motions, events, and faces in the video, the intelligent thumbnail tool can automatically identify, capture, and save highlight frames as video thumbnails.

Directions

1. Add a video to the tool as instructed in steps 1–3 in the Audio/Video format conversion section.
2. After the video is added, the tool will automatically start analyzing the video and capturing highlight frames and display the captured frames on the right.

6. Digital remastering

The digital remastering tool has video noise cancellation, super resolution, SDR to HDR, sharpening, and other capabilities. Through the combination of different capabilities, it can meet your needs for remastering old and low-quality videos.

Directions

1. Add a video to the tool as instructed in steps 1–3 in the Audio/Video format conversion section.
2. After adding the video, select the target resolution and click Start Remastering.

AI processing

1. Image tagging

The image tagging tool can identify scenes, objects, people, and other information in images. It contains thousands of tags in over 60 subcategories in 8 categories, such as natural scenery (mountain, sea, sky, sunset, etc.), man-made environment (building, playground, meeting room, etc.), people (male, female, selfie, group photo, etc.), object (food, clothing, daily necessities, etc.), and pet/wild animal (cat, dog, bird, mammal, marine animal, etc.).

Directions

1. Log in to the COS console and click Smart Toolbox on the left sidebar.
2. On the Smart Toolbox page, select AI Processing > Image Tagging.
3. Add an image to the tool.
4. After the image is added, the tool will automatically start analyzing the image and display image tag information on the right.

2. Vehicle and license plate detection

The vehicle and license plate detection tool can accurately identify the coordinates, brands, models, model years, and colors of almost all passenger cars on the market.

Directions

1. Add an image to the tool as instructed in steps 1–3 in the Image tagging section.
2. After the image is added, the tool will automatically start analyzing the image and display the vehicle and license plate information on the right.

File processing

Online file preview

The file preview tool can convert any Office documents into a webpage format that can be previewed online. After conversion, you can copy the file link to view the file in a browser.

Directions

1. Log in to the COS console and click Smart Toolbox on the left sidebar.
2. On the Smart Toolbox page, select File Preview > Online File Preview.
3. Add a file (e.g., a PPT file) to the tool.
4. After the PPT file is added, the tool will automatically start processing it to generate the HTML webpage for online preview. You can preview it directly or copy the link to preview it in a browser.

Data Processing Workflow

Overview

Last updated：2024-03-25 15:33:39

Overview

Data workflow is a new data processing service launched by COS. It offers two features: workflow and task.
Workflow: With the workflow feature, you can quickly and flexibly create data processing processes as needed. A workflow is bound to a path of an input bucket. When a file is uploaded to the path, the workflow will be automatically triggered to perform the specified processing operation, with the processing result automatically saved to the specified path of the output bucket.
Task: For files in a bucket, you can create tasks to perform data processing operations.
Template: When using data workflows, you usually need to set a series of parameters, which can be combined through a template. This simplifies your operations and allows you to reuse the configured parameters with no need to enter them repeatedly.
Note:
For audio/video transcoding, audio/video splicing, video frame capturing, and animated image generation, you need to specify a template when creating a task or workflow, which can be either a preset or custom one.
Queue and callback: When you activate the data workflow service, the system will automatically create a user queue for you. When you submit a task, the task will be arranged in the queue first and executed in sequence according to the priority and order of submission. You can also set a callback rule to stay up to date with the task or workflow progress, and the system will send the processing result and status information to the specified address.
The processing operations currently supported by data workflow include:
Operation
Description
Audio/Video transcoding
Converts an audio/video file bitstream. It changes parameters of the source bitstream, such as codec, resolution, and bitrate, to adapt to different devices and network conditions.
TESHD transcoding
Provides transcoding capabilities that make videos smaller and clearer. It integrates a complete set of video processing solutions such as image remastering and enhancement, adaptive content parameter selection, and V265 encoder to deliver a better visual experience with guaranteed low network resource usage.
Professional media format transcoding
Can transcode special formats such as XAVC and ProRes.
Video enhancement
Optimizes the video image quality and enhances the visual effect through a series of features, including details enhancement, color enhancement, and SDR to HDR conversion.
Super-resolution
Reconstructs the details and local features of a video by recognizing its content and contour so as to generate a high-resolution video image through a series of low-resolution video images. It can be used in combination with video enhancement to remaster old videos.
Highlight generation
Accurately extracts highlights from a video by recognizing and aggregating the video content, postures, and scenes and quickly clipping them professionally.
Voice/Sound separation
Separates the human voice and background sound in a specified video (or audio) to generate audio materials for subsequent artistic processing of other styles.
Adaptive HLS muxing
Generates a multi-bitrate adaptive file from a raw video to adapt the video to different devices and network conditions.
HDR to SDR conversion
Makes the image details of the output video as close as possible to those of the original video to adapt to different types of devices and avoid image distortion and darkness.
Video frame capturing
Captures the frames of a video at specified time points. You can customize the start time point of frame capturing, frame capturing interval, number of frames to be captured, and output image size and format to meet your diversified needs.
Audio/Video splicing
Adds a video/audio segment at the beginning or end of a video/audio file to generate a new one.
Audio/Video segmentation
Divides a video/audio file into several segments of the specified duration.
Animated image generation
Converts a video file into an animated image file. You can specify the video segment for conversion, frame sampling method, as well as the frame rate, size, and format of the output animated image to meet your different needs.
Intelligent thumbnail generation
Intelligently analyzes the quality, brilliance, and content relevance of video frames by understanding the video content with Tencent Cloud's advanced AI technologies. Then, it extracts optimal frames to generate thumbnails to make the content more engaging.
Speech recognition
Recognizes an audio recording file and returns the recognized text asynchronously. Currently, the supported languages include Mandarin, English, and Cantonese. Meanwhile, Cloud Infinite (CI) can help you process the recognition result, such as blocking impolite words, filtering interjections, and intelligently converting numbers to words, meeting your various speech recognition needs.
File preview
Allows you to preview files of nearly 30 types through image or HTML online, with the source file style preserved as much as possible. This addresses the lack of support for certain file formats on different devices and enables online file preview on PC, app, and other terminals.
Image processing
Supports flexible image editing, such as rotation, cropping, transcoding, and zooming. It provides multiple image downsizing solutions like Guetzli compression, TPG transcoding, and HEIF transcoding, as well as diversified copyright protection solutions like image/text/blind watermarking. This meets your image processing needs in different business scenarios.

Use Cases

Multi-device adaptability

As content platforms are generally intended for multiple types of devices, they need to provide media files in different formats for different users. The audio/video transcoding feature covers most transcoding needs and provides diversified compression capabilities to increase the compression efficiency and downsize files. This reduces lags, storage space usage, and traffic fees.

Video platform

For traditional video platforms, reviewers need to watch videos and then manually select thumbnails, which is labor consuming and slows down video release.
The intelligent thumbnail generation feature can quickly select the most striking frames as thumbnails, which saves labor resources and accelerates video release.
The animated image generation feature allows you to select the highlights in a video to convert them into an animated image for video preview, so that users can get a glimpse of the video without playing it back. Compared with traditional static video thumbnails, animated image thumbnails increase the click rate and video playbacks.

Video subtitles generation

A list of words and corresponding timestamps can be generated for audio files, which makes it easier to subtitle corresponding videos.

Meeting recording transcription

The minutes of large meetings are complex. If a meeting is lengthy and attended by many people, it would be more difficult to keep the complete minutes. The speech recognition feature of COS can recognize Mandarin, English, and Cantonese, which reduces the workload of taking meeting minutes and improves the meeting effect.

Online education

The file preview feature enables viewing various types of files such as courseware and handouts in online education, which delivers an easier user experience.

Website transcoding

The display of file content at websites is subject to browser rules. The file preview feature of COS allows generating images from multiple types of files for preview. This addresses the display problems of file content on webpages.

Directions

Through COS console

You can use the data workflow service in the COS console. For more information, see the documentation for workflow and task configuration.

Through RESTful APIs

You can use APIs to manage workflows and tasks. For more information, see the data workflow API documentation.

Custom Function Processing

Last updated：2024-06-24 16:14:43

Overview

If the existing services or features of COS media processing cannot meet your needs, you can use the custom function processing feature of SCF to write core code logic in order to flexibly implement your business needs while reducing your development costs. For more information on SCF, see Overview.
Note:
Currently, the custom function processing feature can only be initiated in a workflow.
Using SCF for custom processing will incur fees, which will be charged by SCF. For billing details, see Billing Overview.

Directions

1. Log in to the COS console.
2. Select Bucket List on the left sidebar.
3. Click the name of the desired bucket.
4. On the left sidebar, click Data Processing Workflow > Workflow to enter the workflow management page.
5. Click Create Workflow.
6. Click + to the right of “Configure Workflow” to add a custom function node.

7. In the pop-up window, configure the following information.
Input Parameters: The input parameters specified with a custom function in a workflow don't need to be added manually. Instead, they can be obtained according to the node before the custom function.
Namespace: The created function is in the COS namespace by default.
Type: Select Common feature to quickly perform common operations on COS objects or Custom to configure other feature parameters.
Feature: The preset feature is supported when select the common feature.
Function: Currently, only functions that are executed asynchronously and have status tracking enabled are supported in a workflow. To create a function, click Create Function and configure the function as prompted.
8. After confirming that the configuration is correct, click OK.

Configuring a Workflow

Last updated：2025-11-20 15:27:40

Overview

With a data processing workflow, you can quickly and flexibly create video processing processes as needed. A workflow is bound to a path of an input bucket. When a video file is uploaded to the path, the media workflow will be automatically triggered to perform the specified processing operation, with the processing result automatically saved to the specified path of the destination bucket.
You can use a data processing workflow to implement the following features: audio/video transcoding (including top speed codec transcoding and broadcast media format transcoding), video frame capturing, video-to-animated image conversion, intelligent thumbnail, audio/video splicing, voice separation (also known as voice/sound separation), highlights generation (also known as video montage), adaptive multi-bitrate, SDR to HDR, video enhancement, super resolution, audio/video segmentation, custom function, and image processing.
Note:
Currently, workflows can process 3GP, ASF, AVI, DV, FLV, F4V, M3U8, M4V, MKV, MOV, MP4, MPG, MPEG, MTS, OGG, RM, RMVB, SWF, VOB, WMV, WEBM, MP3, AAC, FLAC, AMR, M4A, WMA, and WAV files. When initiating a media processing request, you must enter the complete file name and extension; otherwise, the format cannot be recognized and processed.
Currently, the workflow feature can only manipulate video files being uploaded. To perform media operations on cloud data, use the job feature.

Directions

Creating workflow

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket for media processing.
4. On the left sidebar, select Data Processing Workflow > Workflow to go to the workflow management page.
5. Click Create Workflow.
6. On the Create Workflow page, configure the following items:


Workflow Name: It is required and can contain up to 128 letters, digits, underscores (_), and hyphens (-).
Input Bucket Name: It is the current bucket by default.
Input Path: It is optional and must start and end with /. If it is not specified, the workflow will be triggered for all paths in the input bucket. After the media workflow is enabled, when a video file is uploaded to this path, the workflow will be automatically triggered.
Format: Select the default audio, video, or image file filter rule or a custom rule. You can also select all files to process all objects in the bucket.
Queue: It is required. When you activate the service, the system will automatically create a user queue for you. When you submit a job, the job will be arranged in the queue first and executed in sequence according to the priority and order of submission. You can view the queue information in Common Configuration.
Callback: You can use the queue callback, i.e., callback URL bound to the queue. To modify it, please go to the corresponding queue list.
Configure Workflow: Click + on the right to add any of the following nodes: audio/video transcoding (including top speed codec transcoding and broadcast media format transcoding), video frame capturing, video to animated image conversion, intelligent thumbnail, audio/video splicing, voice separation, highlights generation, HLS adaptive multi-bitrate, SDR to HDR, video enhancement, super-resolution, audio/video segmentation, custom function, and image processing. You need to configure at least one job node in a workflow and set the destination bucket, filename (see Workflow Variable Description), path, and job template for each job node. For more information on templates and how to set them, see Template.
Audio/Video transcoding



Video frame capturing



Video-to-animated image conversion



Intelligent thumbnail



Description: The intelligent thumbnail feature understands the video content with the aid of Tencent Cloud's advanced AI technologies to intelligently extract three optimal keyframes.
Audio/Video splicing



Voice separation



Highlights generation



Adaptive bitrate streaming


Description: The HLS adaptive multi-bitrate feature encapsulates multiple files with multiple bitrates and audio tracks into one multi-bitrate adaptive HLS or DASH video file.
SDRtoHDR



7. After confirming that the configuration is correct, click Save.
Workflows are disabled by default. To enable a workflow, click the toggle in the Enable column. Once enabled, the workflow will take effect in five minutes. Then, it will automatically perform media processing operations on video files uploaded subsequently. After processing files, it will output the new generated files to the specified file path.

Managing workflow

You can view the list of created workflows on the workflow management page.
The workflow list displays the names, IDs, input paths, creation times, and statuses of workflows. You can search for workflows by name and ID to view, edit, or delete specified workflows.
Enable: Once a workflow is enabled, video files uploaded to the specified path in the input bucket will be automatically processed according to the workflow configuration. You can click the toggle again to pause the workflow.
Note:
Workflows are disabled by default. To enable a workflow, click the toggle in the Enable column. Once enabled, the workflow will take effect in 5 minutes.
Details: You can view the configuration details of the current workflow.
View Execution Instance: You can view the workflow execution status and time by time.
More:
Click More > Edit in the Operation column to enter the Edit Workflow page, where you can modify the workflow configuration.
Click More > Delete in the Operation column to delete the workflow.
Note:
You cannot edit or delete an enabled workflow.

Viewing execution instance

An execution instance will be generated after a workflow is executed for each video file. The execution instance page displays the source file address, workflow execution status, and execution time.
1. Go to the workflow management page and click View Execution Instance in the Operation column of the target workflow to enter the execution instance list page.
2. On the list page, click Details in the Operation column of the target instance to enter the instance details page.
3. On the instance details page, you can view the job ID, execution status, start time, and end time of each workflow node.

Triggering workflow

After a workflow is created, it can be automatically triggered for files uploaded to the specified bucket or manually triggered for existing files in the bucket.
1. On the workflow management page, click More > Create Execution Instance of the target workflow.
2. On the Create Execution Instance page, select the file for which to trigger the workflow and click Save to immediately trigger and execute the workflow. You can view the workflow execution status on the execution instance page.




Workflow Variable Description

Workflows support rendering destination file names and URLs with the following variables:
Variable Name
Description
InputName
Filename of the input file (without file extension)
InputNameAndExt
Filename of the input file (with file extension)
InputPath
File input path
RunId
Execution instance ID
Ext
Destination file format
Number
Destination file number

Sample

If the names of your input files are test1.mp4 and test2.mp4, and you want to convert them to the FLV format (the final filenames will be test1.flv and test2.flv), then set the parameter format of the destination filename to ${InputName}.${Ext}.
If the parameter format of the destination filename is set to ${InputNameAndExt}_${RunId}.${Ext}:
When the workflow generates two instances (000001 and 000002) during execution, the final filenames will be test1.mp4_000001.flv and test2.mp4_000002.flv.

Configuring Job

Last updated：2025-11-20 12:01:18

Overview

For files already in a bucket, you can create a job for media processing, speech recognition, file processing, and other operations. Currently, the following jobs are supported: audio/video transcoding, top speed codec transcoding, broadcast media format transcoding, highlights generation (also known as video montage), voice separation (also known as voice/sound separation), audio/video splicing, video enhancement, audio/video segmentation, super resolution, SDR to HDR, video frame capturing, video-to-animated image conversion, intelligent thumbnail, digital watermark extraction, image processing, text to speech, speech recognition, and file preview, some of which can be created by template. You can use the preset templates or customize templates. For more information, see Template.
Note:
Currently, jobs can process 3GP, ASF, AVI, DV, FLV, F4V, M3U8, M4V, MKV, MOV, MP4, MPG, MPEG, MTS, OGG, RM, RMVB, SWF, VOB, WMV, WEBM, MP3, AAC, FLAC, AMR, M4A, WMA, and WAV files. When initiating a media processing request, you must enter the complete file name and extension; otherwise, the format cannot be recognized and processed.
Currently, the job feature can only manipulate existing files. To manipulate files during upload, use the workflow feature as described in Configuring Workflow.
After a job is created, feature fees will be charged by CI. For billing details, see Media Processing Fees.

Viewing Job

On the job page, you can view all jobs in different types for the specified time period, click Job Status to filter and view jobs in different statuses, and search for jobs by job ID in the search box.
1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Click View on the right of a job to view its information:
Job information: Job ID, job status, queue ID, template ID, job creation time, and job end time.
Input information: Source file bucket, region, and storage path.
Output information: Output file address, bucket, region, and storage path.
Note:
A job has six statuses: succeeded, failed, executing, pending, paused, and canceled.
You can query the records of jobs for the past month only.

Creating Audio/Video Transcoding Job

The audio/video transcoding feature converts an audio/video file bitstream. It changes parameters of the source bitstream, such as codec, resolution, and bitrate, to adapt to different devices and network conditions.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Media Processing tab, select Transcoding > Audio/Video Transcoding as the job type, click Create Job, and configure as follows:


Source File URL: Enter the path of the source file, which cannot begin or end with /.
Transcoding Type: Select Standard.
Template Type: Select preset or custom template.
Template: Select the specified template.
Digital Watermark: Add a blind watermark as needed for copyright protection.
Watermark: Add a visible image or text watermark as needed.
Remove Watermark: Remove the watermark.
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Path of the output file.
Destination Filename: Name of the output file.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating Top Speed Codec Transcoding Job

The top speed codec technology improves the subjective image quality of a video at the minimum bitrate. Compared with standard transcoding, it makes videos smaller and clearer and delivers a better visual experience with low network resource usage.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Media Processing tab, select Transcoding > Audio/Video Transcoding as the job type, click Create Job, and configure as follows:
Source File URL: Enter the path of the source file, which cannot begin or end with /.
Transcoding Type: Select Top Speed Codec Transcoding.
Template: Select the specified template.
Digital Watermark: Add a blind watermark as needed for copyright protection.
Watermark: Add a visible image or text watermark as needed.
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Path of the output file.
Destination Filename: Name of the output file.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating Broadcast Media Format Transcoding

This feature produces videos in broadcast media formats such as Apple ProRes and Sony XAVC.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Media Processing tab, select Transcoding > Audio/Video Transcoding as the job type, click Create Job, and configure as follows:
Source File URL: Enter the path of the source file, which cannot begin or end with /.
Transcoding Type: Select Broadcast Media Format.
Template: Select the specified template.
Digital Watermark: Add a blind watermark as needed for copyright protection.
Watermark: Add a visible image or text watermark as needed.
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Path of the output file.
Destination Filename: Name of the output file.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating Highlights Generation Job

The highlights generation feature accurately extracts highlight segments from a video and outputs them as a new file for use in different scenarios subsequently, such as replay and preview.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Media Processing tab, select Smart Editing > Highlights Generation as the job type, click Create Job, and configure as follows:


Source File URL: Enter the path of the source file, which cannot begin or end with /.
Template: Select the specified template.
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Path of the output file.
Destination Filename: Name of the output file.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating Voice Separation Job

The voice separation feature separates the voice from the background sound in a video material to generate a new independent audio file. Then, you can apply artistic processing of other styles to the material without accompaniment and noise.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Media Processing tab, select Smart Editing > Voice Separation as the job type, click Create Job, and configure as follows:


Source File URL: Enter the path of the source file, which cannot begin or end with /.
Template: Select the specified template.
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Path of the output file.
Voice Filename: Name of the output voice file.
Background Sound Filename: Name of the output background sound file.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating Text-to-Speech Job

The text to speech feature can convert text into natural-sounding and smooth speeches for use in smart customer service and audiobook scenarios.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Media Processing tab, select Smart Editing > Text to Speech as the job type, click Create Job, and configure as follows:
Source File URL: Enter the path of the source file, which cannot begin or end with /.
Template: Select the specified template.
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Path of the output file.
Destination Filename: Name of the output audio file.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating Audio/Video Splicing Job

The video/audio splicing feature adds the specified video/audio segment at the beginning or end of a video/audio file to generate a new one.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Media Processing tab, select Transcoding > Audio/Video Splicing as the job type, click Create Job, and configure as follows:


Source File URL: Enter the path of the source file, which cannot begin or end with /.
Template: Select a created audio/video splicing template.
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Storage path of the output file.
Destination Filename: Name of the output file.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating Audio/Video Segmentation Job

The audio/video segmentation feature splits the specified audio/video file into several segments and outputs them in the specified container format.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Media Processing tab, select Transcoding > Audio/Video Segmentation as the job type, click Create Job, and configure as follows:


Source File URL: Enter the path of the source file, which cannot begin or end with /.
Container Format: Select the container format for the output segment.
Segment Duration: Specify the duration of the output segment.
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Storage path of the output file.
Destination Filename: Name of the output file.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating Video Frame Capturing Job

Video frame capturing is a screenshot feature provided by CI to capture the frames of a video at specified time points. After the job is enabled in the console, the output screenshots are in JPG format by default. If you enable captured frame compression, screenshots can be output in HEIF or TPG format.
Note:
A video frame capturing job can be created by template. You can customize the frame capturing start time, frame capturing interval, captured frames, output image size, and output format (captured frame compression needs to be enabled for this option) in a custom video frame capturing template.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Media Processing tab, select Transcoding > Video Frame Capturing as the job type, click Create Job, and configure as follows:


Source File URL: Enter the path of the source file, which cannot begin or end with /.
Template Type: You can select preset or custom template. For more information, see Template.
Template: Select the specified template.
Output: If the video frame capturing job is enabled in the console, screenshots in JPG format will be output by default. If captured frame compression is enabled in the template, screenshots in HEIF or TPG format can be output. If you use the video frame capturing API, you can choose to output JPG or PNG screenshots. For more information, see Getting Media File Screenshot.
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Storage path of the video screenshots.
Destination Filename: Name of the output file. Note that as more than one files are output by smart video frame capturing, the output filename must contain the ${Number} parameter as the sequence number of the screenshot. For example, if the destination file path is set to test-${Number}.jpg and the job captures two screenshots, the actual names of the output files will be test-0.jpg and test-1.jpg.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating Video Enhancement Job

Video enhancement is a video image quality improvement feature provided by CI. You can use it to enhance and beautify image colors and improve the image details.
Note:
A video enhancement job can be created by template. You can customize the color and detail enhancement settings in a custom video enhancement template.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Media Processing tab, select Image Quality Optimization > Video Enhancement as the job type, click Create Job, and configure as follows:


Note:
The input video must be shorter than 30 minutes.
Source File URL: Enter the path of the source file, which cannot begin or end with /.
Enhancement Template: Select a video enhancement template as needed.
Transcoding Template Type: You can select preset or custom template. For more information, see Template.
Transcoding Template: You can select a transcoding template and specify parameters such as resolution, bitrate, and format of the output file.
Digital Watermark: Add a blind watermark as needed for copyright protection.
Watermark: Add a visible image or text watermark as needed.
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Path of the output video.
Destination Filename: Name of the output file.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating Super Resolution Job

The super resolution feature reconstructs the details and local features of a video by recognizing its content and contour so as to generate a high-resolution video image through a series of low-resolution video images. It can be used in combination with video enhancement to remaster old videos.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Media Processing tab, select Image Quality Optimization > Super Resolution as the job type, click Create Job, and configure as follows:
Source File URL: Enter the path of the source file, which cannot begin or end with /.
Super Resolution Template: Select the destination resolution template as needed.
Transcoding Template Type: You can select preset or custom template. For more information, see Template.
Transcoding Template: You can select a transcoding template and specify parameters such as bitrate and format of the output file.
Digital Watermark: Add a blind watermark as needed for copyright protection.
Watermark: Add a visible image or text watermark as needed.
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Path of the output video.
Destination Filename: Name of the output file.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating SDR-to-HDR Job

SDR to HDR is a video dynamic range conversion feature provided by CI. You can use it to convert a standard dynamic range (SDR) video to a high dynamic range (HDR) video.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Media Processing tab, select Image Quality Optimization > SDR to HDR as the job type, click Create Job, and configure as follows:


Note:
The input video must be shorter than 30 minutes.
Source File URL: Enter the path of the source file, which cannot begin or end with /.
HDR Standard: Select HLG or HDR10.
Transcoding Template: Select an H.265 transcoding template. If there are no templates, create an audio/video transcoding template and select H.265 as the encoding format. For more information on how to create a template and configure parameters, see Custom Template.
Watermark: Add a visible image or text watermark as needed.
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Storage path of the destination file after the SDR-to-HDR conversion is completed.
Destination Filename: Name of the output file.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating Video-to-Animated Image Conversion Job

You can use the video-to-animated image conversion feature to convert a video to animated images.
Note:
A video-to-animated image conversion job can be created by template. You can customize the transcoding start time, transcoding duration, frame extraction method, output animated image frame rate, and output animated image size in a custom video to animated image conversion template.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Media Processing tab, select Transcoding > Video-to-Animated Image Conversion as the job type, click Create Job, and configure as follows:


Source File URL: Enter the path of the source file, which cannot begin or end with /.
Template Type: You can select preset or custom template. For more information, see Template.
Template: Select the specified template.
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Storage path of the animated images.
Destination Filename: Name of the output file.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating Intelligent Thumbnail Job

The intelligent thumbnail feature intelligently analyzes the quality, brilliance, and content relevance of video frames by understanding the video content with Tencent Media Lab's advanced AI technologies. Then, it extracts optimal frames to generate thumbnails to make the content more engaging.
Note:
The intelligent thumbnail feature is a paid service and billed by the original video duration. For billing details, see Media Processing Fees.
Three optimal keyframes will be output through smart analysis of each video file.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Media Processing tab, select Smart Editing > Intelligent Thumbnail as the job type, click Create Job, and configure as follows:


Source File URL: Enter the path of the source file, which cannot begin or end with /.
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Storage path of the smart thumbnails.
Destination Filename: Name of the output file.
Note:
As more than one files are output by intelligent thumbnail, the output filename must contain the parameter ${Number} as the thumbnail serial number. For example, if the output file path is set to test-${Number}.jpg, the actual names of the output files will be test-0.jpg and test-1.jpg.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating Digital Watermark Extraction Job

You can use the media processing service to extract the digital watermark from a watermarked video.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Media Processing tab, select Copyright Protection > Digital Watermark Extraction as the job type, click Create Job, and configure as follows:


Source File: Enter the path of the source file, which must begin with but cannot end with /. Different folders are separated with /.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queue.

Creating Speech Recognition Job

The speech recognition feature recognizes a recording file and asynchronously returns the recognized text. It can be used for call center speech quality inspection, video subtitles generation, and meeting recording transcription.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Speech Recognition tab, click Create Job, and configure as follows:
Source File URL: Enter the path of the source file, which cannot begin or end with /.
Recognition Engine: Select a speech recognition engine. Different engines are as described below:
8k_zh: Applies to telephone recording, 8 kHz audio sample rate, Mandarin.
8k_zh_s: Applies to telephone recording, 8 kHz audio sample rate, Mandarin; supports audio separation by speaker.
16k_zh: Applies to audio/video live streaming and video conferences, 16 kHz audio sample rate, Mandarin.
16k_zh_video: Applies to audio/video live streaming, 16 kHz audio sample rate, Mandarin.
16k_en: Applies to English audio, 16 kHz audio sample rate.
16k_ca: Applies to Cantonese audio, 16 kHz audio sample rate.
Sound Channels: Select mono-channel or dual-channel.
Recognition Result: Speech recognition result text output by sentence or word (only supported for the Chinese speech recognition engines at 16 kHZ audio sample rate).
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Storage path of the recognized text.
Destination Filename: Name of the output file.
Filter Restricted Words: Select whether to filter restricted words or replace them with \*.
Filter Modal: Select whether to filter modal.
Smart Speech Conversion: After it is enabled, recognized Chinese numbers will be converted to Arabic numbers.
Queue: Currently, only the default speech recognition queue queue-speech-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating File Preview Job

The file preview feature allows you to preview files of nearly 30 types online through image or HTML, with the source file style preserved as much as possible. This addresses the lack of support for certain file formats on different devices and enables easy online file preview on PC, app, and other terminals.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the File Preview tab, click Create Job, and configure as follows:
Source File URL: It cannot begin or end with /; for example, doc/example.doc.
Preview Setting: Select to preview whole document or specified page. A job supports up to 5,000 pages. If more pages are input, only the first 5,000 pages can be converted.
Output Format: Currently, JPG and PNG formats are supported for output images. The PDF format is supported only for whole document preview.
Destination Bucket: Select a bucket for which the file preview feature has been enabled in the current region.
Destination Path: It is optional. If it is not set, it will be the same as the input file path.
Destination Filename: The file preview service converts each page of the original file into an image. Therefore, you need to add a placeholder (${Number} or ${Page}) to the output filename to number the output images. The output numbers are the same as the file page numbers. For example, if you want to preview a file with three pages and set the output filename to output${Number}.jpg, then three images output1.jpg, output2.jpg, output3.jpg will be output.
Queue: Currently, only the default file preview queue queue-doc-process-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Creating Image Processing Job

The image processing feature supports flexible image editing, such as rotation, cropping, transcoding, and scaling. It provides multiple image downsizing solutions like Guetzli compression, TPG transcoding, and HEIF transcoding, as well as diversified copyright protection solutions like image/text/blind watermarking. This meets your image processing needs in different business scenarios.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to manipulate.
4. On the left sidebar, select Data Processing Workflow and click Job to enter the job management page.
5. Select the Image Processing tab, click Create Job, and configure as follows:
Input Bucket Name: It is the current bucket by default.
File Path: Enter the path of the source file, which must begin with but cannot end with /. Different folders are separated with /.
Template: Select the specified template.
Destination Bucket: Select a bucket for which the media processing feature has been enabled in the current region.
Destination Path: Storage path of the image processing result.
Destination Filename: Name of the output file.
Queue: Currently, only the default queue queue-1 is supported. For more information, see Queues and Callbacks.
Queue Callback URL: Callback URL bound to the queue. You can configure it in the queue in Common Configuration.

Template

Last updated：2024-06-24 16:14:43

Overview

When using the data processing workflow feature, you usually need to set a series of parameters, which can be combined through a template. This simplifies your operations and allows you to reuse the configured parameters with no need to enter them repeatedly.
For media processing features such as audio/video transcoding, audio/video splicing, video frame capturing, and video to animated image conversion, you need to specify a template when creating a job or workflow. The template page provides system templates, and you can also customize templates based on your business needs.

System Templates

The system combines common parameters in advance into system templates, so that you can use them directly. When creating a job or workflow, you can select such a template by the template name.
1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket that you want to operate.
4. On the left sidebar, select Data Processing Workflow > Template Configuration to view templates of different processing types.

Click View in the Operation column to view template details.
Note:
Currently, the system provides 15 audio/video transcoding, 3 video frame capturing, and 18 video to animated image conversion templates.
You can view the information of system templates but cannot edit or delete them.

System templates for audio/video transcoding

Template ID
Template Name
Container Format
Video Encoding Format
Resolution
Video Bitrate
Audio Encoding Format
Audio Bitrate
t0e2b9f4cd25184c6ab73d0c85a6ee9cb5
H264-MP4-LD-360P
MP4
H.264
640 * proportionally scaled
512 Kbps
AAC
128 Kbps
t0876739cd865042d1957d73c78f0484fb
H264-MP4-SD-480P
MP4
H.264
720 * proportionally scaled
1024 Kbps
AAC
128 Kbps
t0852e7ff4acd4484e99ba104f3840d3cb
H264-MP4-HD-720P
MP4
H.264
1280 * proportionally scaled
2000 Kbps
AAC
128 Kbps
t04df9eb0c373c4a8780ec894ce05469a7
H264-MP4-FHD-1080P
MP4
H.264
1920 * proportionally scaled
3500 Kbps
AAC
128 Kbps
t09d027135634d47048e5a30dc1e19ee90
H264-MP4-2K
MP4
H.264
2048 * proportionally scaled
4800 Kbps
AAC
128 Kbps
t0e634622e8dfb49339ba478d60ddc7188
H264-HLS-LD-360P
M3U8
H.264
640 * proportionally scaled
512 Kbps
AAC
128 Kbps
t0fa5bdbf58bb348e88bf73fae5d674fdf
H264-HLS-SD-480P
M3U8
H.264
720 * proportionally scaled
1024 Kbps
AAC
128 Kbps
t09d0f419921e44ed98190f355ec9fd629
H264-HLS-HD-720P
M3U8
H.264
1280 * proportionally scaled
2000 Kbps
AAC
128 Kbps
t080ae8a06f9074f3daa46201078f8d4b1
H264-HLS-FHD-1080P
M3U8
H.264
1920 * proportionally scaled
3500 Kbps
AAC
128 Kbps
t0ab68939cef0f40d19c4a135df540239f
H264-HLS-2K
M3U8
H.264
2048 * proportionally scaled
4800 Kbps
AAC
128 Kbps
t0e165bef65ed24d568eeecc8661248af6
H264-FLV-LD-360P
M3U8
H.264
640 * proportionally scaled
512 Kbps
AAC
128 Kbps
t057d0410c32444e48b9220f9571e6097a
H264-FLV-SD-480P
M3U8
H.264
720 * proportionally scaled
1024 Kbps
AAC
128 Kbps
t00daf332ba39049f8bfb899c1ed0134b0
H264-FLV-HD-720P
M3U8
H.264
1280 * proportionally scaled
2000 Kbps
AAC
128 Kbps
t0d41905a814434c8a81897ecb54d53a32
H264-FLV-FHD-1080P
M3U8
H.264
1920 * proportionally scaled
3500 Kbps
AAC
128 Kbps
t0e287e59454b94a8983ba78a6a30ee864
H264-FLV-2K
M3U8
H.264
2048 * proportionally scaled
4800 Kbps
AAC
128 Kbps

System templates for video frame capturing

Template ID
Template Name
Frame Capturing Start Time
Frame Capturing Interval
Max Frame Count Per Video
Output Image Size
Output Format
t01d40e440761448fc8c538fb8d5a5b81e
snapshot_320 * 180_1
0s
2s
5
320 * 180 px
JPEG
t0a60a2bc71a4b40c7b3d7f7e8a2779a81
snapshot_640 * 360_2
0s
10s
5
640 * 360 px
JPEG
t07740e32081b44ad7a0aea03adcffd54a
snapshot_1280 * 720_3
0s
10s
5
1280 * 720 px
JPEG

System templates for video to animated image conversion

Template ID
Template Name
Transcoding Start Time
Transcoding Duration
Frame Extraction Method
Output Animated Image Frame Rate
Output Animated Image Size
Output Format
Output Animated Image Quality
t04373959a69c04d47b62fd214dd13d8e9
gif_320 * 180_1
0s
600s
Only keyframes are extracted
Adaptive
320 * 180 px
GIF
\
t0341b0ab2b8a340ff826e9cb4f3a7baea
gif_320 * 180_2
0s
600s
One frame is extracted every 10s
Adaptive (0.1 FPS)
320 * 180 px
GIF
\
t046b1d8e5bdf842c6a58d8028b48eafee
gif_320 * 180_3
0s
600s
Ten frames are extracted per second
Adaptive (10 FPS)
320 * 180 px
GIF
\
t0ef2077f215864c018a2fca73614ceca6
gif_640 * 360_4
0s
600s
Only keyframes are extracted
Adaptive
640 * 360 px
GIF
\
t0d21406ca737a40869973a37a5daa349a
gif_640 * 360_5
0s
600s
One frame is extracted every 10s
Adaptive (0.1 FPS)
640 * 360 px
GIF
\
t0878a9c9c1f054cb5bca68b8b06e556c2
gif_640 * 360_6
0s
600s
Ten frames are extracted per second
Adaptive (10 FPS)
640 * 360 px
GIF
\
t0dae821708cea4ba5b3e271810ac80a21
gif_1280 * 720_7
0s
600s
Only keyframes are extracted
Adaptive
1280 * 720 px
GIF
\
t03fef67ad94d2466b9c0c89252ed72e87
gif_1280 * 720_8
0s
600s
One frame is extracted every 10s
Adaptive (0.1 FPS)
1280 * 720 px
GIF
\
t030a64e9f9f5a4f53a9ef64bb7ce490b5
gif_1280 * 720_9
0s
600s
Ten frames are extracted per second
Adaptive (10 FPS)
1280 * 720 px
GIF
\
t03b0e9eca4fc34e2cba9da89d9c7c13a2
webp_320 * 180_1
0s
60s
Only keyframes are extracted
Adaptive
320 * 180 px
WEBP
75
t016fcddf6bc3c44b793e9b7b07119b4ee
webp_320 * 180_2
0s
600s
One frame is extracted every 10s
Adaptive (0.1 FPS)
320 * 180 px
WEBP
75
t0bf1f1ce6d2404b258c0f81fbb9aaece1
webp_320 * 180_3
0s
600s
One frame is extracted every 10s
Adaptive (10 FPS)
320 * 180 px
WEBP
75
t098d6d3fcfd2c45309a408594a42559f6
webp_640 * 360_4
0s
60s
Only keyframes are extracted
Adaptive
640 * 360 px
WEBP
75
t0169a6a9c2eec4b51972eb63bafcbf08d
webp_640 * 360_5
0s
600s
One frame is extracted every 10s
Adaptive (0.1 FPS)
640 * 360 px
WEBP
75
t0ef9ba537011e4876b8777aebc19d10a5
webp_640 * 360_6
0s
600s
One frame is extracted every 10s
Adaptive (10 FPS)
640 * 360 px
WEBP
75
t02743d344b5e74c579e50e9e135b432b8
webp_1280 * 720_7
0s
60s
Only keyframes are extracted
Adaptive
1280 * 720 px
WEBP
75
t0dd27c136ff2741538bec96981e058868
webp_1280 * 720_8
0s
600s
One frame is extracted every 10s
Adaptive (0.1 FPS)
1280 * 720 px
WEBP
75
t00ad05235d67a45a9a697b553052b7346
webp_1280 * 720_9
0s
600s
One frame is extracted every 10s
Adaptive (10 FPS)
1280 * 720 px
WEBP
75

Custom Template

If system templates cannot meet your needs, use custom templates. Currently, you can create custom templates for audio/video transcoding, top speed codec transcoding, highlights generation, video frame capturing, video to animated image conversion, video watermark, audio/video splicing, voice/sound separation, video enhancement, super-resolution, image processing, and broadcast media format transcoding.

Audio/Video transcoding

The audio/video transcoding feature converts an audio/video file bitstream. It changes parameters of the source bitstream, such as codec, resolution, and bitrate, to adapt to different devices and network conditions. You can customize the template parameters in a custom audio/video transcoding template.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket where you want to store the video.
4. On the left sidebar, select Data Processing Workflow > Template Configuration to enter the template configuration page.
5. Select Media Processing > Audio/Video Transcoding and click Create Transcoding Template.

6. In the Create Transcoding Template window, configure the following items:
Template Name: It can contain up to 64 letters, digits, underscores (_), hyphens (-), and asterisks (*).
Transcoding Type: You can select video or audio transcoding.
Container Format: Video transcoding supports the MP4, FLV, HLS, TS, MKV, and WebM formats. Audio transcoding supports the MP3, AAC, AMR, FLAC, and WebM formats.
Encoding Format: H264, H265, AV1, VP8, and VP9 are supported.
Bitrate: Adaptive bitrate is supported, which can automatically analyze the video content to set the optimal bitrate.
Transcoding Duration: You can select the input file duration or customize the duration.
Audio/Video Parameters: You can customize audio/video parameters as needed.
7. Click OK. After successfully creating the template, you can view, edit, or delete it in the custom template list.
Note:
You can use the template when creating an audio/video transcoding job as instructed in Configuring Job or creating a workflow as instructed in Configuring Workflow.

TSC transcoding

The top speed codec transcoding feature improves the subjective image quality of a video at a low bitrate. Compared with regular audio/video transcoding, it outputs smaller files and clearer video images and delivers a better visual experience with guaranteed low network resource usage. You can customize parameters such as codec, resolution, and bitrate in a custom top speed codec transcoding template.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket where you want to store the video.
4. On the left sidebar, select Data Processing Workflow > Template Configuration to enter the template configuration page.
5. Select Top Speed Codec Transcoding and click Create Top Speed Codec Transcoding Template.
6. In the Create Broadcast Media Format Transcoding Template window, configure the following items:
Template Name: It can contain up to 64 letters, digits, underscores (_), hyphens (-), and asterisks (*).
Transcoding Type: It is video transcoding by default.
Container Format: Supported formats include MP4 and HLS.
Transcoding Duration: You can select the input file duration or customize the duration.
Audio/Video Parameters: You can customize audio/video parameters as needed.
7. Click OK. After successfully creating the template, you can view, edit, or delete it in the custom template list.
Note:
You can use the template when creating a broadcast media format transcoding job as instructed in Configuring Job or creating a workflow as instructed in Configuring Workflow.

Broadcast media format transcoding

This feature processes special formats such as XAVC and ProRes.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket where you want to store the video.
4. On the left sidebar, select Data Processing Workflow > Template Configuration to enter the template configuration page.
5. Select Broadcast Media Format Transcoding and click Create Broadcast Media Format Transcoding Template.
6. In the Create Broadcast Media Format Transcoding Template window, configure the following items:
Template Name: It can contain up to 64 letters, digits, underscores (_), hyphens (-), and asterisks (*).
Preset Encoder Configuration: Select the default values of encoder parameters such as sample rate.
Container Format: Supported formats include MXF.
Transcoding Duration: You can select the input file duration or customize the duration.
Audio/Video Parameters: You can customize audio/video parameters as needed.
7. Click OK. After successfully creating the template, you can view, edit, or delete it in the custom template list.
Note:
You can use the template when creating a broadcast media format transcoding job as instructed in Configuring Job or creating a workflow as instructed in Configuring Workflow.

Highlights generation

The highlights generation feature automatically extracts highlights from a video. You can use a custom template to set the highlights generation template name and specify the maximum duration, resolution, and format of the output highlight video.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket where you want to store the video.
4. On the left sidebar, select Data Processing Workflow > Template Configuration to enter the template configuration page.
5. Select Highlights Generation and click Create Highlights Generation Template.
6. In the Create Voice Separation Template window, configure the following items:
Note:
Currently, highlights generation can be used only for landscape, food, street, and vlog scenarios and will support more scenarios in the future. If you want to customize this feature, contact us for assistance.
Template Name: It can contain up to 64 letters, digits, underscores (_), hyphens (-), and asterisks (*).
Container Format: Supported formats include MP4, FLV, HLS, TS, and MKV.
Highlight Video Duration: You can select the duration of the complete output highlight video after automatic analysis or customize the duration.
Audio/Video Parameters: You can customize audio/video parameters as needed.
7. Click OK. After successfully creating the template, you can view, edit, or delete it in the custom template list.
Note:
You can use the template when creating a highlights generation job as instructed in Configuring Job or creating a workflow as instructed in Configuring Workflow.

Video frame capturing

The video frame capturing feature captures the frames of a video at specified time points. The output screenshots are in JPG format by default. If you enable captured frame compression, screenshots can be output in HEIF or TPG format. You can customize the template name, frame capturing start time, frame capturing interval, captured frames, and output image size and format in a custom video frame capturing template.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket where you want to store the video.
4. On the left sidebar, select Data Processing Workflow > Template Configuration to enter the template configuration page.
5. Select Video Frame Capturing and click Create Video Frame Capturing Template.

6. In the Create Video Frame Capturing Template window, configure the following items:
Template Name: It can contain up to 64 letters, digits, underscores (_), hyphens (-), and asterisks (*).
Frame Capturing Start Time: You can select any time point within the full video length.
Frame Capturing Method:
All frames will be captured by default: Every video frame will be captured.
Custom Frame Capturing Interval: Frames will be captured at specified time intervals from the frame capturing start time to the end of the video.
Even Frame Capturing: Frames will be captured at even intervals calculated based on the specified number of captured frames from the frame capturing start time to the end of the video.
Capture Keyframes: Capture only keyframes.
Max Frame Count Per Video: This parameter is required if you select All frames will be captured by default, Custom Frame Capturing Interval, or Capture Keyframes as the frame capturing method.
Captured Frames: This parameter is required if you select Even Frame Capturing as the frame capturing method. Frames will be captured at even intervals calculated based on the specified number of frames captured from the frame capturing start time to the end of the video.
Output Image Size: The default output screenshot size is the same as that of the original video image. If you select custom image size, you must enter an integer between 128 and 4096 for the width and height respectively.
Video Frame Compression: If it is enabled, captured images can be compressed.
7. Click OK. After successfully creating the template, you can view, edit, or delete it in the custom template list.
Note:
You can use the template when creating a video frame capturing job as instructed in Configuring Job or creating a workflow as instructed in Configuring Workflow.

Video-to-animated image conversion

The video to animated image conversion feature converts a video to animated images. You can customize the template name, transcoding start time, transcoding duration, frame extraction method, output animated image frame rate, and output animated image size in a custom video to animated image conversion template.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket where you want to store the video.
4. On the left sidebar, select Data Processing Workflow > Common Configuration. Then, select the Template tab to go to the template configuration page.
5. Select Video to Animated Image Conversion and click Create Video to Animated Image Conversion Template.

6. In the Create Video to Animated Image Conversion Template window, configure the following items:
Template Name: It can contain up to 64 letters, digits, underscores (_), hyphens (-), and asterisks (*).
Transcoding Start Time: You can select any time point within the full video length.
Transcoding Duration: It specifies the duration of video transcoding after the transcoding start time. You can select Original Video Duration or Custom Duration.
Frame Extraction Method:
Extract all frames: Every video frame will be extracted.
Frame Extraction Frequency: You can set the number of frames to be extracted per second (an integer between 1 and 50).
Frame Extraction Interval: Frame will be extracted at the specified intervals in seconds.
Extract key frames only: The system will intelligently identify and extract the optimal set of frames based on AI understanding of the video content and output them as an animated image.
Output Animated Image Frame Rate: If Adaptive is selected, the system will automatically select an appropriate frame rate based on the settings of the above parameters. You can also select Custom Playback Frame Rate to restrict the frame rate to 1–60 FPS.
Output Animated Image Format: The output animated image is in GIF format by default. If you select the WEBP format, you need to select the animated image quality, which ranges from 1 to 99 and is 75 by default.
Output Animated Image Size: The default output animated image size is the same as that of the original video. If you select custom width and height, you must enter an integer between 128 and 4096 for the width and height respectively.
7. Click OK. After successfully creating the template, you can view, edit, or delete it in the custom template list.
Note:
You can use the template when creating a video-to-animated image conversion job as instructed in Configuring Job or creating a workflow as instructed in Configuring Workflow.

Video watermark

The video watermark feature adds a text or image watermark to a video during transcoding.
Note:
Currently, you can add up to three watermarks in the console or five via API at a time. To add more watermarks, contact us for assistance.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket where you want to store the video.
4. On the left sidebar, select Data Processing Workflow > Template Configuration to enter the template configuration page.
5. Select Video Watermark and click Create Video Watermark Template.

6. In the Create Video Watermark Template window, configure the following items:
Common parameters
Template Name: It can contain up to 64 letters, digits, underscores (_), hyphens (-), and asterisks (*).
Watermark Type: Select image or text watermark.
Origin Point: Select the top-left, top-right, bottom-left, or bottom-right corner.
Opacity: Select a value between 1% and 100%.
Offset Method: The watermark offset is based on the origin point. You can select offset by ratio or fixed position.
Watermark Duration: Select Same as video duration or Specified period. If you select the latter, you can set the watermark start time and end time. If you set the start time only, the watermark will be displayed until the video ends by default.
Image watermark parameters
Select Image: If you select image watermark, you need to select its source. Currently, only a watermark image in the same bucket can be selected. If the bucket doesn't have desired images, you need to upload a new one.
Image Layer: Select whether to place the image on top of or underneath the video.
If the image is placed on top of the video, the effect is as shown below:
img

If the image is placed underneath the video (as the background), the effect is as shown below:
img

Watermark Dimensions:
Input image size: The original watermark image size will be retained without any processing. Note that if the watermark image is larger than the video image, the watermark cannot be completely displayed.
By ratio: You can set the percentage (1–100) of only the width or height or both of them. If the width or height is not set, it will be scaled proportionally. Suppose the width ratio is a and height ratio is b, then the watermark width will be w = W * a, and the watermark height will be h = H * b (here, W and H are the video width and height respectively).
Fixed size: You can specify the watermark width and height between 8 and 4096 px.
Offset Method (On top of video):
By ratio: You can set the percentage (0–100) of the width or height. As shown below, suppose the horizontal offset ratio is a and the vertical ratio is b, then the horizontal offset will be Dx = W * a, and the vertical offset will be Dy = H * b (here, W and H are the video width and height respectively).
Fixed position: Select a value between 0 and 4096 px. The horizontal offset is Dx, and the vertical offset is Dy.
Watermark Dimensions:
Input image size: The original watermark image size will be retained without any processing. Note that if the watermark image is smaller than the video image, the watermark cannot be completely displayed.
By ratio: You can set the percentage (100–300) of only the width or height or both of them. If the width or height is not set, it will be scaled proportionally. Suppose the width ratio is a and height ratio is b, then the watermark width will be w = W * a, and the watermark height will be h = H * b (here, W and H are the video width and height respectively).
Fixed size: You can specify the watermark width and height between 8 and 4096 px.
Offset Method (Underneath video):
By ratio: You can set the percentage (-300–0) of the width or height. As shown below, suppose the horizontal offset ratio is a and the vertical ratio is b, then the horizontal offset will be Dx = W * a, and the vertical offset will be Dy = H * b (here, W and H are the video width and height respectively).
Fixed position: Select a value between -4096 and 0 px. The horizontal offset is Dx, and the vertical offset is Dy.
Text watermark parameters
Watermark Text: It can contain up to 64 letters, digits, underscores (_), hyphens (-), and asterisks (*).
Font Size: Select a value between 5 and 100 px.
Font: Select Ariblk, Arial, Ahronbd, Helvetica, or HelveticaNeue.
Font Color: It is in the format of 0xRRGGBB.
7. Click OK.
After successfully creating the template, you can view, edit, or delete it in the custom template list.
You can click Preview to view the position and dimensions of the watermark in videos of three common resolutions and quickly adjust the template.
Note:
You can use the template when [creating an audio/video transcoding, SDR-to-HDR, video enhancement, or super resolution job as instructed in Configuring Job or creating a workflow as instructed in Configuring Workflow.

Audio/Video splicing

The video/audio splicing feature adds the specified video/audio segment at the beginning or end of a video/audio file to generate a new one.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket where you want to store the video.
4. On the left sidebar, select Data Processing Workflow > Common Configuration. Then, select the Template tab to go to the template configuration page.
5. Select Audio/Video Splicing and click Create Audio/Video Splicing Template.
6. In the Create Audio/Video Splicing Template window, configure the following items:
Template Name: It can contain up to 64 letters, digits, underscores (_), hyphens (-), and asterisks (*).
Container Format: Supported formats include AAC, MP3, MP4, FLV, HLS, and TS.
Splicing Position: Select whether to add the file at the beginning or end of the source file.
Other parameters: You can customize the audio/video parameters as needed.
7. Click OK. After successfully creating the template, you can view, edit, or delete it in the custom template list.
Note:
You can use the template when creating an audio/video splicing job as instructed in Configuring Job or creating a workflow as instructed in Configuring Workflow.

Voice separation

You can separate the same audio file into a voice file and a background sound file for subsequent video editing and playback.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket where you want to store the video.
4. On the left sidebar, select Data Processing Workflow > Template Configuration to enter the template configuration page.
5. Select Smart Audio > Voice Separation and click Create Voice Separation Template.

6. In the Create Voice Separation Template window, configure the following items:
Template Name: It can contain up to 64 letters, digits, underscores (_), hyphens (-), and asterisks (*).
Output Audio Format: Supported formats include MP3, AAC, AMR, and FLAC.
Output Audio: Specify to output voice or background sound.
Sample Rate: Select an option as needed.
Audio Bitrate: Enter a value as needed.
Channels: Select an option as needed.
7. Click OK. After successfully creating the template, you can view, edit, or delete it in the custom template list.
Note:
You can use the template when creating a voice separation job as instructed in Configuring Job or creating a workflow as instructed in Configuring Workflow.

Text to speech

You can convert text to speeches in different voices for use in audiobook, navigation, and other scenarios.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket where you want to store the video.
4. On the left sidebar, select Data Processing Workflow > Template Configuration to enter the template configuration page.
5. Select Smart Audio > Text to Speech and click Create Text-to-Speech Template.
6. In the Create Text-to-Speech Template window, configure the following items:
Template Name: It can contain up to 64 letters, digits, underscores (_), hyphens (-), and asterisks (*).
Output Audio Format: Supported formats include MP3, AAC, AMR, and FLAC.
Voice: Specify the output voice.
Processing Mode: Select async or sync processing.
Volume: Adjust the output volume.
Speech Speed: Adjust the output speech speed.
7. Click OK. After successfully creating the template, you can view, edit, or delete it in the custom template list.
Note:
You can use the template when creating a text-to-speech job as instructed in Configuring Job or creating a workflow as instructed in Configuring Workflow.

Image quality enhancement

Video enhancement uses AI to improve the video quality and enhance the video colors and details visually. It includes super resolution, detail enhancement, SDR-to-HDR, and frame interpolation features.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket where you want to store the video.
4. On the left sidebar, select Data Processing Workflow > Template Configuration to enter the template configuration page.
5. Select Image Quality Enhancement and click Create Image Quality Enhancement Template.
6. In the Create Image Quality Enhancement Template window, configure the following items:
Note:
Currently, video enhancement supports color enhancement, detail enhancement, super resolution, and SDR-to-HDR. Other features will be provided in the future.
The input video for enhancement must be shorter than 30 minutes.
Template Name: It can contain up to 64 letters, digits, underscores (_), hyphens (-), and asterisks (*).
Color Enhancement: Select automatic system analysis for color enhancement or customize the color enhancement parameters.
Detail Enhancement: Select automatic system analysis for detail enhancement or customize the detail enhancement parameters.
7. Click OK. After successfully creating the template, you can view, edit, or delete it in the custom template list.
Note:
You can use the template when creating a video enhancement job as instructed in Configuring Job or creating a workflow as instructed in Configuring Workflow.

Image processing

Image processing is a rich-featured, cost-effective, and high-reliability image processing service provided by CI. It supports flexible image editing, such as rotation, cropping, transcoding, and scaling. It provides multiple image downsizing solutions like Guetzli compression, TPG transcoding, and HEIF transcoding, as well as diversified copyright protection solutions like image/text/blind watermarking. This meets your image processing needs in different business scenarios.

Directions

1. Log in to the COS console.
2. Click Bucket List on the left sidebar.
3. Click the name of the bucket where you want to store the video.
4. On the left sidebar, select Data Processing Workflow > Common Configuration. Then, select the Template tab to go to the template configuration page.
5. Select Video Processing and click Create Video Processing Template.
6. In the Create Video Processing Template window, configure the following items:
Template Name: It can contain up to 64 letters, digits, underscores (_), hyphens (-), and asterisks (*).
Editing Mode: Select Basic or Enhanced. The latter delivers greater image reconstruction and restoration effects.
Basic Processing: Select the output target resolution.
Text Watermark: After it is enabled, you can add a single or tiled watermark to the image.
Image Watermark: After it is enabled, you can add an animated or static watermark at the specified position on the image.
Preview: You can preview the processing effect.
7. Click OK. After successfully creating the template, you can view, edit, or delete it in the custom template list.
Note:
You can use the template when creating an image processing job as instructed in Configuring Job or creating a workflow as instructed in Configuring Workflow.

Queues and Callbacks

Last updated：2025-11-20 12:02:41

Overview

When you activate the data processing workflow service, the system will automatically create a user queue for you. When you submit a job, the job will be arranged in the queue first and executed in sequence according to the priority and order of submission. You can also set a callback rule to stay up to date with the job or workflow progress, and the system will send the processing result and status information to the specified address. The queues for different services are as follows:
Feature Name
Queue Name
Media processing
queue-1
Speech recognition
queue-speech-1
File preview
queue-doc-process-1
Note:
Currently, one feature supports only one queue. If you want more concurrent queues, contact us.

Directions

Enabling or pausing queue

You can enable or pause a queue in its Operation column.
1. Log in to the COS console.
2. Select Bucket List on the left sidebar.
3. Click the name of the bucket for video storage.
4. On the left sidebar, select Data Processing Workflow > Common Configuration. Then, click Queues and Callbacks to enter the queue configuration page to enable or pause a queue.
Note:
After a queue is paused, its jobs will stop, and you cannot use the job and workflow features in the console.

Setting callback rule

COS supports user-defined callback URLs. After an event is completed, the system sends an HTTP POST request whose body contains notification content to a user-defined callback URL. You can use the configured callback URL to learn about the processing progress and status so that you can perform other operations as needed.
1. Log in to the COS console.
2. Select Bucket List on the left sidebar.
3. Click the name of the bucket for video storage.
4. On the left sidebar, select Data Processing Workflow > Common Configuration. Then, click Queues and Callbacks to enter the queue configuration page.
5. Click Configure Callback Rule in the Operation column of the target queue.
6. In the pop-up window, set the status to enable or disable callback.


7. When enabling callback, you need to specify a URL for the system to send HTTP requests. For more information on callback, see Callback content.

Callback content

After a job is completed, the system will send the following callback content to the configured callback URL:
<Response>
      <JobsDetail></JobsDetail>
      <NonExistJobIds></NonExistJobIds>
</Response>
The parameters are described as follows:
Parameter
Description
Type
JobsDetail
Job details. Same as Response.JobsDetail in CreateMediaJobs.
Container
NonExistJobIds
List of non-existing job IDs queried. If all jobs exist, this parameter is not returned.
String


Application Integration

Add COS Inventory Analysis Function

Last updated：2025-05-27 10:51:52

Overview

COS Inventory Analysis offers inventory analysis services for various scenarios, helping you extract key information from inventory files more efficiently. This article introduces how to add a COS Inventory Analysis function. After creating the function, please refer to Setting Inventory Analysis to use the COS Inventory Analysis feature.

Must-Knows

If you have added a COS inventory analysis rule to your bucket via the COS console, you can see the COS inventory analysis function you created in the SCF console. Do not delete this COS inventory analysis function. Otherwise, your rule may not take effect.
The current feature is only supported in regions where both COS and SCF are the same. It is not supported in some regions, such as Chongqing. The specific regions are subject to those displayed on the COS and SCF consoles.
COS inventory analysis feature depends on cloud function service. The cloud function service provides users with a free quota. The part exceeding the free quota is charged according to SCF product pricing.

Directions

1. Log in to the COS Console.
2. In the left sidebar, select Application Integration > Extended Function, and find COS inventory analysis.
3. Click Configure Analysis Feature to enter the function rule configuration page for COS Inventory Analysis.
4. On the page, select the region where you need to add a function, then click Add Function, and configure the following information in the pop-up window:
Function Name: The function name serves as the unique identifier of the function and cannot be modified after creation. You can view the function on the SCF console.
Associated Bucket: Select a COS storage bucket with inventory analysis feature enabled.
Execution Method: Currently only supports async execution. After calling, it enters the async execution status, does not return the execution result, but supports a longer running time.
Authentication Method: Currently only support SCF authentication.
Status: To use this feature, the bucket needs to have the inventory feature configured. For operation guide, please see Setup Checklist.
SCF Authorization: COS inventory analysis requires authorization for the cloud function to read the corresponding file from your bucket and save the inventory analysis result file to a user-specified location. Therefore, this item needs to be selected to complete the authorization of the SCF service.
5. After confirming that the configuration is correct, click Confirm. Wait a moment, and you will see that the function has been added successfully. According to the pop-up notification, you can go to the inventory configuration item of the bucket to create an inventory analysis task. For details, refer to COS Inventory Analysis.
6. You can perform the following operations on the newly created function:
Click Instructions to reopen the use guidance and quickly go to the COS storage bucket.
Click Log to view the historical running status of the COS inventory analysis function. When an error occurs, you can also click Log to quickly jump to the SCF console to view the error details in the log.
Click Details to view detailed information about the function and trigger.
Click More > Edit to view the COS inventory analysis function rules.
Click More > Delete to delete the non-use COS inventory analysis function rule.

Setting COS Inventory Analysis

Last updated：2025-08-15 16:46:17

Overview

After you add the inventory feature of Cloud Object Storage (COS), you can enable the inventory analysis feature to perform further data analysis on the generated inventory file. The inventory analysis feature will integrate the inventory data within a specified time period for statistical analysis and extract key indicators for you to view.

Prerequisites

Use the inventory analysis feature. First, you need to add the COS inventory service. Then, create an inventory analysis serverless cloud function (SCF). For the creation guide, please refer to add COS inventory analysis function.
Note:
The inventory analysis feature does not support instant inventory. Please use scheduled inventory.

Directions

1. Log in to the COS Console.
2. In the left sidebar, click Bucket List to enter the bucket list management page.
3. Find the bucket that needs inventory analysis. Click the Bucket Name to go to the bucket management page.
4. In the left sidebar, select Basic Configurations > Inventory.
Note:
Before using the analysis feature of the inventory, ensure that the inventory has been added and the inventory file has been generated.
5. In the inventory analysis configuration item, if you haven't added a COS inventory analysis function, click Use Now below. At this point, the system will detect whether you have added COS inventory analysis function rules. See add COS inventory analysis function to add the function.
If you have added a COS inventory analysis function, select the corresponding function name in the Inventory Analysis area, select the time period of the inventory you want to analyze, click New Analysis Task, and configure the following information in the pop-up window:
Note:
The total number of objects in the bucket for inventory analysis must not exceed 100 million.
Basic Configuration



Inventory Name: Select the inventory to be analyzed.
Inventory timestamp: The value of the timestamp in the object key of the inventory file.
Product Destination Directory: The result file and manifest file of this analysis are both sent to this directory.
Job Description: Customize the description for this analysis.
Filter Configuration



Filter Condition: Support filtering based on object key, object size, latest modification time, and whether it is the latest version.
Object Key: Enter the object key template. Wildcards are supported *, for example:test/* represents objects with the test/ prefix.
Latest modified: Supports yesterday, the last 3 days, or a custom time range.
Object Size: Ranges from the size of specified objects.
Latest Version: Support selection of whether it is the latest version.
Statistics Configuration



Job Type: Option to export statistics or export object list.
Statistics dimension: Optional directory prefix, object suffix, storage type.
Statistics content: Optional size of object, number of objects.
Statistics entry limit: Please enter the upper limit of statistics entries, which is the maximum number of rows of statistical results in the result file. It must not exceed 10,000 rows.
6. After confirming that the configuration is correct, click OK, and you will see that the analysis task has been added successfully.
You can perform the following operations on the created analysis task:
Click View result to view the results of this inventory analysis and the save location of the result files.
Note:
You can query inventory analysis tasks created in the last 3 days.
You can view the result only after an inventory analysis task is completed. An inventory analysis usually takes several to dozens of minutes depending on the inventory size.
Click Running Log to navigate to the SCF console and view the running log of the COS inventory analysis.

MySQL Backup

Last updated：2024-10-24 18:41:53

Overview

MySQL Backup is a SCF-based feature provided by COS. It allows you to store data in TencentDB for MySQL to COS so that data can be stored persistently and protected from data loss or corruption. After you set a backup function rule for a bucket, SCF will scan your MySQL backup files periodically and store them in the bucket.

Notes

MySQL Backup functions only back up the backup files of TencentDB for MySQL. Therefore, if MySQL backup has not been enabled, the functions cannot be executed. For more information about TencentDB for MySQL backups, see Backing up Databases.
If you have added a MySQL backup rule to your bucket in the COS console, the backup function will appear in the SCF console. DO NOT delete the function. Otherwise, your rule may not take effect.
MySQL Backup is supported in SCF-enabled regions, including Guangzhou, Shanghai, Beijing, Chengdu, Hong Kong (China), Singapore and Silicon Valley. For more supported regions, see SCF Documentation.

Directions

Setting backup via Application Integration

1. Log in to the COS console.
2. On the left sidebar, click Application Integration > Data Backup and find MySQL Backup.
3. Click Configure Backup Rule to go to the rule configuration page.
4. Click Add Function.
Note:
If you haven't activated SCF, go to the SCF console to activate it and authorize the service as instructed.
5. In the pop-up window, configure the following items:
Function Name: Uniquely identifies a function and cannot be modified after being set. You can view the function in the SCF console.
Associated Bucket: A bucket to store the MySQL backup files
Note:
The MySQL data to be backed up must be in the same region as the associated bucket.
Trigger Period: Triggers the backup operation for the MySQL Backup function. Every day, every week, and custom periods are supported.
Cron Expression: If you use a custom period, you can use Cron to specify the trigger period rule. Cron follows the local Standard Time. For detailed configuration policies, see Timer Trigger Description.
Database Instance: MySQL instance in the region where the current bucket resides
Delivery Path: Delivery path prefix of the backups. If not specified, backups will be stored in the root directory of the bucket.
SCF Authorization: SCF needs to be authorized so that it can read the MySQL instances as well as their backup files, and store the backup files to the specified bucket.
6. Click Confirm. You can perform the following operations on the created function:
Click Log to view the historical running status of MySQL backup. If an error is reported, you can click Log to quickly redirect to the SCF console for viewing the error log details.
Click More > Edit to modify the MySQL data backup rule.
Click More > Delete to delete the unwanted MySQL backup rule.

Setting backup via bucket configuration

1. Log in to the COS console.
2. Click Bucket List on the left sidebar and then click the desired bucket for the backup.
3. Click Function Service > MySQL Backup Function.
Note:
If you haven't activated SCF, go to the SCF console to activate it and authorize the service as instructed.
4. Click Add Function.
5. In the pop-up window, configure the following items:
Function Name: Uniquely identifies a function and cannot be modified after being set. You can view the function in the SCF console.
Trigger Period: Triggers the backup operation for the MySQL Backup function. Every day, every week, and custom periods are supported.
Cron Expression: If you use a custom period, you can use Cron to specify the trigger period rule. Cron follows China Standard Time (UTC+8:00). For detailed configuration policies, see Cron Documentation.
Database Instance: MySQL instance in the region where the current bucket resides
Delivery Path: Delivery path prefix of the backups. If not specified, backups will be stored in the root directory of the bucket.
SCF Authorization: SCF needs to be authorized so that it can read the MySQL instances as well as their backup files, and store the backup files to the specified bucket.
6. Click Confirm. You can perform the following operations on the created function:
Click Log to view the historical running status of MySQL backup. If an error is reported, you can click Log to quickly redirect to the SCF console for viewing the error log details.
Click More > Edit to modify the MySQL data backup rule.
Click More > Delete to delete the unwanted MySQL backup rule.

MongoDB Backup

Last updated：2025-07-10 14:54:18

Overview

MongoDB Backup is a SCF-based feature provided by COS. It allows you to back up data in TencentDB for MongoDB to COS so that data can be stored persistently and protected from data loss or corruption. After you set a backup function rule for a bucket, SCF will scan your MongoDB backup files periodically and store them in the bucket.

Notes

MongoDB Backup functions only back up the backup files of TencentDB for MongoDB. Therefore, if MongoDB backup has not been enabled, the functions cannot be executed. For more information about TencentDB for MongoDB backups, see Data Backup.
If you have added a MongoDB backup rule to your bucket in the COS console, the backup function will appear in the SCF console. DO NOT delete this function. Otherwise, your rule may not take effect.
MongoDB Backup is supported in SCF-enabled regions, including Guangzhou, Shanghai, Nanjing, Beijing, Chengdu, Hong Kong (China), Singapore, Bangkok, Seoul, Tokyo, Silicon Valley and Virginia. For more supported regions, see SCF Documentation.
The size of a single MongoDB backup file cannot exceed 300GB.

Directions

1. Log in to the COS console.
2. On the left sidebar, click Application Integration > Data Backup and find MongoDB Backup.
3. Click Configure Backup Rule to go to the rule configuration page.
4. Click Add Function.
Note:
If you haven't activated SCF, go to the SCF console to activate it and authorize the service as instructed.
5. In the pop-up window, configure the following items:
Function Name: Uniquely identifies a function and cannot be modified after being set. You can view the function in the SCF console.
Associated Bucket: A bucket to store the MongoDB backup files.
Trigger Period: Triggers the backup operation for the MongoDB Backup function. Every day and custom periods are supported.
Note:
For example, a user configures a MongoDB backup rule on July 3, 2023, specifying that the MongoDB instance files will be backed up to COS daily at 00:00.
The SCF will approximately start running on July 4, 2023, at 00:00:03 (there may be a few seconds delay in SCF startup), scanning MongoDB instance files from the past 3 days and uploading them to COS.
As the SCF execution time is 2023-07-04 00:00:03, the final backup file time range is: 2023-07-01 00:00:00 ~ 2023-07-04 00:00:03.
Cron Expression: If you use a custom period, you can use Cron to specify the trigger period rule. Cron is currently running in UTC +8 China Standard Time (CST), which is Beijing Time. For detailed configuration policies, see Timer Trigger Description.
Database Instance: MongoDB instance in the region where the current bucket resides
Delivery Path: Delivery path prefix of the backups. If not specified, backups will be stored in the root directory of the bucket.
SCF Authorization: SCF needs to be authorized so that it can read the MongoDB instances as well as their backup files, and store the backup files to the specified bucket.
6. Click Confirm. You can perform the following operations on the created function:
Click Log to view the historical running status of MongoDB backup. If an error is reported, you can click Log to quickly redirect to the SCF console for viewing the error log details.
Click Details to view detailed information about the function and trigger.
Click More > Edit to modify the MongoDB data backup rule.
Click More > Trigger to immediately trigger the SCF without waiting for the specified period.
Click More > Delete to delete the unwanted MongoDB backup rule.

SQL Server Data Backup

Last updated：2025-07-10 15:20:32

Overview

SQL Server data backup is a SCF-based database backup feature provided by COS. It allows you to dump SQL Server data to COS so that data can be stored persistently and protected from data loss or corruption. After you set a backup function rule for a specified bucket, SCF will scan your server backup files periodically and dump them in the bucket.

Notes

SQL Server data backup functions back up only the backup files of SQL Server. If SQL Server backup is not enabled, the backup functions cannot be executed. For more information about Tencent Cloud SQL Server database backup, please refer to Backup Cloud Database SQL Server.
If you have added a SQL Server data backup rule to your bucket in the COS console, you can view it in the SCF console. DO NOT delete the function. Otherwise, your rule may not take effect.
The current feature is only supported in regions where both COS and SCF are the same. It is not supported in some regions, such as Chongqing. The specific regions are subject to those displayed on the COS and SCF consoles.

Directions

1. Log in to the COS console.
2. On the left sidebar, click Application Integration > Data Backup.
3. Find SQL Server Backup and click Configure Backup Rule to go to the rule configuration page.
4. Click Add Function.
Note:
If you haven't activated SCF, go to the SCF console to activate it and authorize the service as instructed.
5. In the pop-up window, configure the following items:
Function Name
Uniquely identifies a function and cannot be modified after being set. You can view the function in the SCF console.
Associated Bucket
A bucket to store SQL Server backup files.
Trigger Period
Triggers the backup operation for the SQL Server backup function. Every day and custom periods are supported.
Cron Expression
If you use a custom period, you can use Cron to specify the trigger period rule. Cron follows the local Standard Time. For detailed configuration policies, see Timer Trigger Description.
Database Instance
SQL Server instance list of the region where the current bucket resides
Delivery Path
Delivery path prefix of the backups. If not specified, backups will be stored in the root directory of the bucket.
SCF Authorization
SCF needs to be authorized so that it can read the SQL Server instances as well as their backup files, and save the backup files to the specified bucket.
6. Click Confirm. You can perform the following operations on the created function:
Click Log to view the historical running status of the SQL Server data backup. If an error is reported, you can click Log to quickly redirect to the SCF console for viewing the error log details.
Click Details to view detailed information about the function and trigger.
Click More > Edit to modify the SQL Server data backup rule.
Click More > Trigger to manually trigger the SQL Server backup rule.
Click More > Delete to delete the unwanted SQL Server data backup rule.

CKafka Message Backup

Last updated：2024-10-24 18:41:53

Overview

CKafka message backup is provided by COS based on SCF to dump CKafka messages to COS, which facilitates data analysis and download.
Based on the open-source Apache Kafka message queuing engine, Tencent Cloud Kafka (CKafka) provides high-throughput and highly scalable message queuing services. For more information, see Overview.
If a backup rule has been configured for a bucket, messages generated in a CKafka instance will be dumped to the COS bucket according to the specified time granularity.

Notes

If you have added a CKafka message backup rule to your bucket via the COS console, the function will appear in the SCF console. DO NOT delete the function. Otherwise, your rule may not take effect.
CKafka message backup is supported in SCF-enabled regions, including Guangzhou, Hong Kong (China), Shanghai, Beijing, Chengdu, Singapore and Silicon Valley. For more supported regions, see the SCF documentation.

Directions

1. Log in to the COS console.
2. On the left sidebar, click Application Integration > Data Backup and find CKafka Message Backup.
3. Click Configure Backup Rule to go to the rule configuration page.
4. Click Add Function.
Note:
If you haven't activated SCF, go to the SCF console to activate it and authorize the service as instructed.
5. In the pop-up window, configure the following items:
Function Name: Uniquely identifies a function and cannot be modified after being set. You can view the function in the SCF console.
Associated Bucket: A COS bucket that stores CKafka messages
Time Granularity: You can set an interval (5−15 minutes) to aggregate messages according to the message volume. The dumping performance is affected by the number of files aggregated, the number of partitions, and the value of partition_max. For more information, see Partition in Glossary.
SCF Authorization: (required) SCF needs to be authorized to read messages of CKafka instances and dump the messages to the specified bucket.
6. Click Next to configure CKafka as follows:
Instance: A CKafka instance that is used as the message source. Only instances in the same region are supported.
Topic: A topic used as the message source
Start Point: The topic offset for dumping the backup messages
Address: A VPC address is required. CKafka instances that use a classic network require a routing policy. For more information, see Adding Routing Policy.
Note:
An IP address in the corresponding VPC subnet must be available and support DHCP.
7. Click Next to configure delivery. The configuration item is as follows:
Destination Path: A path to deliver the backup messages. If this field is not specified, the files will be stored in the root directory of the bucket. To use a prefix, end it with a slash (/). 8. Click Confirm. You can perform the following operations on the created function:
Click Log to view the historical running status of CKafka message backup. If an error is reported, you can click Log to quickly redirect to the SCF console for viewing the error log details.
Click More > Edit to modify the CKafka message backup rule.
Click More > Delete to delete the unwanted CKafka message backup rule.

TDMQ Message Backup

Last updated：2024-06-24 16:14:43

Overview

TDMQ message backup is provided by COS based on SCF to dump TDMQ messages to COS, which facilitates data analysis and download.
TDMQ is a proprietary finance-grade distributed message middleware developed based on Pulsar, an open-source project of Apache. It features high cross-region consistency, reliability, and concurrency. For more information, see Overview.
If a backup rule has been configured for a bucket, SCF will dump messages generated in TDMQ to a COS bucket according to the specified time granularity.

Notes

If you have added a TDMQ message backup rule to your bucket via the COS console, the function will appear in the SCF console. DO NOT delete the function. Otherwise, your rule may not take effect.
Currently, regions that support backing up TDMQ messages to COS include Guangzhou, Shanghai, Hong Kong (China), Beijing, Chengdu, Singapore, and Silicon Valley.

Directions

1. Log in to the COS console.
2. On the left sidebar, click App Integration > Data Backup and find TDMQ Message Backup.
3. Click Configure Backup Rule to go to the rule configuration page.
4. Click Add Function.
Note:
If you haven't activated SCF, go to the SCF console to activate it and authorize the service as instructed.
5. In the pop-up window, configure the following items:
Function Name: Uniquely identifies a function and cannot be modified after being set. You can view the function in the SCF console.
Associated Bucket: A COS bucket that stores TDMQ messages
Time Granularity: An interval (5−15 minutes) to aggregate messages according to the message volume. Each message file can be up to 500 MB and contain up to 5,000 pieces of messages.
SCF Authorization: (required) SCF needs to be authorized to read TDMQ messages and dump the messages to the specified bucket.
6. Click Next to configure TDMQ as follows:
Cluster: A TDMQ cluster as the message source. Only TDMQ clusters in the same region are supported.
Namespace: A namespace in the cluster
Topic: A topic used as the message source
Subscription: Select a subscription. If existing subscriptions cannot meet your needs, you can create a new one in the TDMQ console.
Start Point: Start point of historical messages
Role: Select the TDMQ role (a TDMQ-specific concept, which is different from that mentioned in Tencent Cloud), which is the minimum unit for permission division within TDMQ. You can create multiple roles and grant them different message production/consumption permissions in different namespaces.
Role Key: Select the TDMQ role key, which is an authentication tool. You can add a key in the client to access TDMQ to produce/consume messages. Each role has a unique key corresponding to itself.
Address: It must be a VPC access address. The TDMQ cluster should be connected to the VPC. For more information, see VPC Access.
Note:
An IP address in the corresponding VPC subnet must be available and support DHCP.
7. Click Next to configure delivery. The configuration item is as follows: Destination Path: A path to deliver the backup messages. If this field is not specified, the files will be stored in the root directory of the bucket. To use a prefix, end it with a slash (/).
8. Click Confirm. You can perform the following operations on the created function:
Click Log to view the historical running status of TDMQ message backup. If an error is reported, you can click Log to quickly redirect to the SCF console for viewing the error log details.
Click More > Edit to modify a TDMQ message backup rule.
Click More > Delete to delete an unwanted TDMQ message backup rule.

Redis Backup

Last updated：2025-10-09 14:59:26

Feature Overview

Redis Backup is a SCF-based feature provided by COS. It allows you to store data in TencentDB for Redis to COS so that data can be stored persistently and protected from data loss or corruption. After you set a backup function rule for a bucket, SCF will scan your Redis backup files periodically and store them in the bucket.

Supports and Limits

Redis Backup functions back up only the backups of TencentDB for Redis. Therefore, if Redis Backup has not been enabled, the functions cannot be executed. For more information about TencentDB for Redis backup, see Backing Up TencentDB for Redis.
If you have previously added a Redis Backup rule to a bucket in the COS console, you can view the created Redis Backup function in the SCF console. Please do not delete this function, as it may cause your rule to become ineffective.
The current feature is only supported in regions where COS and SCF are available. Partial regions (such as Chongqing) are not supported. Refer to the regions displayed in the COS and SCF console for details.

Instructions

1. Log in to the COS Console.
2. In the left sidebar, click Application Integration > Data Backup to enter the application integration management page.
Note:
In addition to the feature entrance, you can also click bucket list, select the corresponding bucket name, click Function Compute > Redis backup function, and enter the function configuration interface to add.
3. Locate Redis Backup and click Configure Backup Rule to access the rule configuration page.
4. Click Add Function.
Note:
If you have not activated the SCF service, please go to the SCF console to activate it and follow the instructions to complete the service authorization.
5. In the pop-up window, configure the following items:
Function Name: Uniquely identifies a function and cannot be modified after being set. You can view the function in the SCF console.
Associated bucket: The bucket storing Redis RDB backup files. If configuring in the Function Compute feature item of the specified bucket, this configuration item is not available.
Trigger cycle: The Redis backup function triggers backup transfer operations via a scheduled trigger, supporting daily or custom periods.
Note:
For example, a user configured a Redis backup rule on July 3, 2023, specifying to backup the Redis instance file to COS at 00:00 daily.
The cloud function starts running around 00:00:03 on July 4, 2023 (with a few seconds delay during startup), scans Redis instance files from the past 3 days, then uploads them to COS.
Since the SCF execution time was 2023-07-04 00:00:03, the backup file time range is: 2023-07-01 00:00:00 ~ 2023-07-04 00:00:03.
Cron expression: When the trigger cycle is set to custom, you can specify the trigger period rule via Cron. Cron currently runs on UTC+8 China Standard Time, which is Beijing time. For detailed configuration policy, please refer to Cron related documentation.
Database Instance: TencentDB for Redis instance in the same region where the current bucket resides.
Delivery Path: Delivery path prefix of the backups. If not specified, backups will be stored in the root directory of the bucket.
SCF Authorization: SCF needs to be authorized so that it can read the Redis instances as well as their backups, and store the backups to the specified bucket.
6. After adding the configuration, click Confirm to see that the function has been added successfully.
7. You can perform the following operations on the newly created function.
Click Log to view the historical running status of Redis backup. If an error is reported, you can click Log to quickly redirect to the SCF console for viewing the error log details.
Click Details to view the function and trigger detailed information.
Click More > Edit to modify the Redis Backup rule.
Click More > Trigger to manually trigger the Redis data backup rule.
Click More > Delete to remove unused Redis Backup rules.

CDN Log Backup

Last updated：2024-10-24 18:41:53

Overview

After a domain name is connected to Content Delivery Network (CDN), users' resource requests will be scheduled to CDN nodes. If a CDN node has the resource cached, it returns the resource directly. Otherwise, the request will be passed through to the origin server to pull the requested resource.
CDN nodes respond to most of the user requests. To facilitate access analysis, CDN packages access logs of the entire network at an hourly granularity.
The CDN log backup feature is provided by COS based on SCF to dump CDN logs to COS, which facilitates access behavior analysis and service quality monitoring.
After you configure a log backup rule for a bucket, SCF will dump the CDN logs to the bucket according to the time granularity configured.

Notes

Once you added a CDN log backup rule to your bucket via the COS console, the backup function can be viewed in the SCF console. DO NOT delete the function. Otherwise, your rule may not take effect.
CDN log backup is supported in SCF-enabled regions, including Guangzhou, Shanghai, Hong Kong (China), Beijing, Chengdu, Singapore and Silicon Valley. For more supported regions, see the SCF documentation.

Directions

1. Log in to the COS console.
2. On the left sidebar, click Application Integration > Data Backup and find CDN Log Backup.
3. Click Configure Backup Rule to go to the rule configuration page.
4. Click Add Function.
Note:
If you haven't activated SCF, go to the SCF console to activate it and authorize the service as instructed.
5. In the pop-up window, configure the following items:


Function Name: Uniquely identifies a function and cannot be modified after being set. You can view the function in the SCF console.
Associated Bucket: A COS bucket that stores the CDN logs
Trigger Period: A period to trigger the CDN log backup (a timer is used). Every day and custom periods are supported.
Cron Expression: If you use a custom period, you can use Cron to specify the trigger period rule. Cron follows the Local Standard Time. For detailed configuration policies, see Timer Trigger Description.
CDN Acceleration Domain: One or multiple domains whose logs are to dump
Destination Path: A path to deliver the logs. You can deliver logs to the root directory or specify a path prefix.
SCF Authorization: (required) SCF needs to be authorized to read CDN logs and dump them to the specified bucket.
6. Click Confirm. After the CDN backup rule is created, you can view it in the list. You can perform the following operations on the created CDN log backup rule:
Click Log to view the historical running status of CDN log backup. If an error is reported, you can click Log to quickly redirect to the SCF console for viewing the error log details.
Click More > Edit to modify the CDN log backup rule.
Click More > Delete to delete the unwanted CDN log backup rule.

CLS Log Backup

Last updated：2024-10-24 18:41:53

Overview

The CLS log backup feature is provided by COS based on SCF to dump CLS logs to COS.
After you configure a log backup rule for a bucket, SCF will dump the CLS logs to the bucket according to the time granularity configured.
CLS log backup can ship log data to COS to further meet the needs of log backup scenarios and harness the value of log data. Log backup is an asynchronous process. When log data is generated, SCF automatically backs up the log data to COS for storage via trigger.

Notes

If you have previously added a CLS log backup rule to your bucket via the COS console, you can view the CLS log backup function you created in the SCF console. DO NOT delete the function. Otherwise, your rule may not take effect. CLS log backup is supported in SCF-enabled regions, including Guangzhou, Shanghai, Beijing, Hong Kong (China), Chengdu, Singapore and Silicon Valley. For more supported regions, see the SCF documentation.

Directions

1. Log in to the COS console.
2. On the left sidebar, click Application Integration > Data Backup and find CLS Log Backup.
3. Click Configure Backup Rule to go to the rule configuration page.
4. Click Add Function.
Note:
If you haven't activated SCF, go to the SCF console to activate it and authorize the service as instructed.
5. In the pop-up window, configure the following items:
Function Name: Uniquely identifies a function and cannot be modified after being set. You can view the function in the SCF console.
Associated Bucket: A COS bucket that stores the CLS logs.
Logset: A logset is a project management unit of CLS, used to distinguish logs of different projects. You can select the logset of the message source, which must be in the function region.
Log Topic: A log topic is the basic management unit of CLS and also the smallest unit for managing and configuring CLS triggers. A logset can contain multiple log topics. You can select a log topic of the message source.
Max Wait Time: You can set this parameter to control the log obtaining frequency. The parameter value can range from 3 to 300 seconds. If you set the parameter to 300 seconds, the SCF will collect the log data generated within 300 seconds and package it centrally as log files for backup.
SCF Authorization: SCF needs to be authorized to read CLS logs and dump them to the specified bucket.
6. Click Next to configure the following information:
Compression Format: You can determine whether to compress log files before backup. A compressed log file can be up to 128 KB. Currently, log files can be compressed using gzip, lzop or snappy.
Partition Format: A directory is automatically generated based on the strftime syntax. For example, if the partition format is %Y/%m/%d/%Y%m%d%H%M, the generated directory is 2021/06/25/202106252232.
Delivery Path: Log backup path. You can select the root directory or specify a path prefix.
Delivery Sample: The final backup filename is in the format of {COS bucket}{Directory prefix}{Partition format}_{random}.{type}.
7. Click Confirm. After the CLS backup rule is created, you can view it in the list.
You can perform the following operations on the created CLS log backup rule:
Click Log to view the historical running status of CLS log backup. If an error is reported, you can click Log to quickly redirect to the SCF console for viewing the error log details.
Click More > Edit to modify the CLS log backup rule.
Click More > Delete to delete the unwanted CLS log backup rule.

Adding Log Analysis Function

Last updated：2025-05-27 10:48:44

Overview

The COS log analysis feature applies to various scenarios to help you efficiently extract key information from log files. This document describes how to add a COS log analysis function. After adding a function, see Setting Log Analysis to use this feature.

Notes

If you have added a COS log analysis rule to your bucket via the COS console, you can view the COS log analysis function you created in the SCF console. Do not delete the COS log analysis function. Otherwise, your rule may not take effect.
The current feature is only supported in regions where both COS and SCF are the same. It is not supported in some regions (for example, Chongqing). The specific regions are subject to those displayed on the COS and SCF consoles.
The log analysis feature depends on the SCF service, which provides users with a free tier. You will be billed for the part exceeding the free tier according to SCF pricing.

Directions

1. Log in to the COS console.
2. On the left sidebar, click App Integration > Extended Features and find COS Log Analysis.
3. Click Configure Analysis Feature to enter the function rule configuration page.
4. Select a region to add the function and click Add Function. Configure the following information in the pop-up window:
Function Name: Uniquely identifies a function and cannot be modified after it is created. You can view the function in the SCF console.
Associated Bucket: Select a COS bucket with the log analysis feature enabled.
Execution Method: Only async execution is supported. After the function is called, tasks will be executed asynchronously without returning the execution result. A longer running time is supported.
Authentication Method: Only SCF is supported.
Log storage status: Need confirmation on whether the log storage status is enabled. For operation guide, please see Set Log Storage.
SCF Authorization: SCF need to be authorized to read files from your bucket and save the result file to the specified folder.
5. After confirming that the configuration is correct, click Confirm. Wait a moment, and you will see that the function has been added. Based on the pop-up notification, you can go to the log management configuration item of the bucket to create a log analysis task. For details, refer to COS Log Analysis.
6. Click Confirm.
You can perform the following operations on the created function:
Click Instructions to reopen the usage guidance and quickly go to the COS storage bucket.
Click Log to view the historical running status.
Click Details to view detailed configuration rules.
Click More > Edit to modify a COS log analysis rule.
Click More > Delete to delete an unwanted COS log analysis rule.

Data Export to CKafka

Last updated：2024-06-24 16:14:43

Overview

Data export to CKafka is a data export solution provided by COS based on SCF. It can help you export data in CSV, JSON, and other formats to the CKafka service in the same region for aggregation and analysis of massive amount of message and log data.

Notes

Data export to CKafka involves COS' data extraction APIs. For more information on the restrictions, see SELECT Overview.
If you have added a rule of data export to CKafka to your bucket in the COS console, the export function will appear in the SCF console. Do not delete or modify the function; otherwise, your rule may not take effect.
Currently, the data export to CKafka feature is only supported in Guangzhou, Shanghai, Beijing, and Chengdu regions.
The data export to CKafka feature depends on the SCF service, which provides a free tier. Excessive usage will be billed at SCF prices. For more information, see Pricing.

Directions

1. Log in to the COS console.
2. On the left sidebar, click App Integration > Data Export and find Data Export to CKafka.
3. Click Configure Rule to enter the rule configuration page.
4. Click Add Function.
Note:
If you haven't activated SCF, go to the SCF console to activate it and authorize the service as instructed.
5. In the pop-up window, configure the following information:

Function name prefix: It uniquely identifies a function and cannot be modified after being set. You can view the function in the SCF console.
Scenario: Select the log source from which you want to export data. We recommend you select COS log file export.
Source bucket: It is the name of the bucket where the logs are stored. If you select COS log file export, you need to enable the COS log storage feature first.
Log storage status: Make sure that the status is Enabled.
Destination Bucket: The bucket to store the logs.
Log delivery prefix: Enter a path prefix that makes it easy for you to find logs.
SCF authorization: This authorization is required, because data export to CKafka requires you to authorize SCF to read logs from your bucket.
6. Click Enter preset parameter to set the data export to CKafka configuration items as follows:

Access method: Select the CKafka access method. If you select VPC, you need to enter the VPC information.
Address: Enter your CKafka access address.
Topic: Enter the name of your CKafka topic.
Authentication configuration: Select your CKafka authentication method. If you select Authentication required, you need to enter the corresponding username and password.
Max rate: Set the rate cap for exporting data to CKafka.
7. If you want to personalize log data extraction, click Previous to configure options. Generally, we recommend you directly click Confirm to add the function.

You can perform the following operations on the created function:
Click Log to view the historical running status of data export to CKafka. If an error is reported, you can click Log to quickly redirect to the SCF console to view its details.
Click Details to view the detailed configuration of the function.
Click Edit to modify a rule of data export to CKafka.
Click Trigger and select an existing log in the bucket to directly trigger the export to CKafka.
Click Delete to delete an unwanted rule of data export to CKafka.

Data Export to ES

Last updated：2024-06-24 16:14:43

Overview

Data export to ES is a data export solution provided by COS based on SCF. It can help you export data in CSV, JSON, and other formats to the ES service in the same region for quick setup of log analysis and exception monitoring use cases.

Notes

Data export to ES involves COS' data extraction APIs. For more information on the restrictions, see SELECT Overview.
If you have added a rule of data export to ES to your bucket in the COS console, the export function will appear in the SCF console. Do not delete or modify the function; otherwise, your rule may not take effect.
Currently, the data export to ES feature is only supported in Guangzhou, Shanghai, Beijing, and Chengdu regions.
The data export to ES feature depends on the SCF service, which provides a free tier. Excessive usage will be billed at SCF prices. For more information, see Pricing.

Directions

1. Log in to the COS console.
2. On the left sidebar, click App Integration > Data Export and find Data Export to ES.
3. Click Configure Rule to enter the rule configuration page.
4. Click Add Function.
Note:
If you haven't activated SCF, go to the SCF console to activate it and authorize the service as instructed.
5. In the pop-up window, configure the following information:

Function name prefix: It uniquely identifies a function and cannot be modified after being set. You can view the function in the SCF console.
Scenario: Select the log source from which you want to export data. We recommend you select COS log file export.
Source bucket: It is the name of the bucket where the logs are stored. If you select COS log file export, you need to enable the COS log storage feature first.
SCF authorization: This authorization is required, because data export to ES requires you to authorize SCF to read logs from your bucket.
6. Click Enter preset parameter to set the data export to ES configuration items as follows:

Service version: Select the ES version, which can be 7.x or 6.x.
Access method: Select the ES access method. If you select VPC, you need to enter the VPC information.
Address: Enter the ES access address, which must start with http:// or https://.
Index name: Enter the name of the ES search index you created.
Authentication configuration: Select your ES authentication method. Currently, you can only select Authentication required for ES, and you need to enter the corresponding username and password.
Max rate: Set the rate cap for exporting data to ES.
7. If you want to personalize log data extraction, click Previous to configure options. Generally, we recommend you directly click Confirm to add the function.
1

You can perform the following operations on the created function:
Click Log to view the historical running status of data export to ES. If an error is reported, you can click Log to quickly redirect to the SCF console to view its details.
Click Details to view the detailed configuration of the function.
Click Edit to modify a rule of data export to ES.
Click Trigger and select an existing log in the bucket to directly trigger the export to ES.
Click Delete to delete an unwanted rule of data export to ES.