Serverless FAQs
Last updated:2025-11-20 16:15:42
This document introduces FAQs related to Serverless.
What Is the Serverless Service?
TDSQL-C for MySQL supports two modes of database instances: provisioned resource and Serverless. The Serverless service is a serverless architecture version of TDSQL-C for MySQL, which charges based on actual computing and storage resource usage and can automatically scale in/out according to business loads. This maximizes resource utilization and optimizes cost control. It is suitable for low-frequency database usage, such as development and testing environments, or scenarios with fluctuating business peaks.
How Is the Serverless Service Billed?
Computing and storage resources in Serverless are billed separately: Computing fees are charged based on the number of CCUs, and storage fees are charged based on the amount of storage used (GB). Billing is calculated per second and settled on an hourly basis. The billing formula is as follows:
Total Serverless fees = Compute node fees + Storage space fees
= Serverless compute unit price x CCU usage + Storage space price x storage capacity
What Is CCU?
TDSQL-C Compute Unit (CCU) is the billing unit for computing resources in Serverless. One CCU is approximately equal to one CPU and 2 GB of memory.
For each billing cycle, the number of used CCUs is determined by taking the maximum value between the number of CPU cores used by the database and half of the allocated memory (in GB).
What Are the Key Features of the Serverless Service?
The Serverless service provides the following key features:
The adjustable elastic scaling range of CCUs allows the Serverless cluster to automatically increase or decrease the number of CCUs within this range based on actual business pressure, making it more adaptable to fluctuating computing demands.
The Serverless service continuously monitors the user's CPU, memory, and other workload conditions and triggers the automatic scaling policy based on predefined rules.
The Serverless service supports custom auto-pause settings. If no connection is detected, the instance automatically pauses and instantly resumes without interruption when a task connection is received.
The kernel team has implemented relevant optimizations, enabling the Serverless service to have elastic anti-jitter capabilities, which make the cluster's elastic scaling more stable.
How Do I Create a Serverless Cluster?
You can create a Serverless cluster through the Purchase page by selecting Serverless as the instance mode. For details, see Creating a Serverless Cluster.
In Which Regions Is the Serverless Cluster Available?
Currently, Serverless is supported only in Guangzhou, Shanghai, Beijing, Chengdu, Nanjing, Hong Kong (China), Silicon Valley, Virginia, Singapore, Tokyo, and Seoul. If your desired region is not listed, you can submit a ticket to provide feedback.
Which Database Versions Are Supported for Serverless Clusters?
Serverless clusters support MySQL 5.7 and MySQL 8.0, the same as those supported by provisioned resource clusters.
What Deployment Architectures Are Available for Serverless Clusters?
Both single-node edition and cluster edition architectures are supported.
Single-node edition architecture: A cluster with one read-write instance. It can be switched to cluster edition architecture in the console after purchase.
Cluster edition architecture: A cluster with one read-write instance and 1-15 read-only instances (optional). It can be switched to single-node edition architecture in the console after purchase.
What Is the Compute Unit Elasticity Range for Serverless Clusters?
The minimum computing power for a single Serverless instance can be set to 0.25 CCU, while the maximum computing power can be set to 64 CCU. Due to different resources supported by different availability zones (AZs), see Service Computing Power Configuration for the specific elastic range of the computing power.
Note:
It is recommended to set the minimum capacity to 1 CCU and select a higher value for the maximum capacity when configuring the elastic range for the first time. A lower capacity setting allows the cluster to scale in to the maximum extent when idle, avoiding unnecessary fees. A higher capacity setting enables the cluster to scale out as needed during peak workloads, ensuring stable performance during business traffic surges.
What Is the Maximum Storage Space for Serverless Clusters?
The maximum storage space varies based on the compute unit configuration. The maximum supported storage space is 384 TB. For details, see Service Compute Unit Configuration.
How Much Cost Savings Can a Serverless Cluster Provide Compared to a Provisioned Resource Cluster?
To estimate the cost savings of switching from a non-Serverless cluster (provisioned resource cluster) to a Serverless cluster, you can use the Serverless Cost Estimator for comparison.
Can a Serverless Cluster Mount a Read-Only Instance from a Provisioned Resource Cluster?
It is not supported.
Can a Provisioned Resource Cluster Mount a Serverless Read-Only Instance?
Yes, this can be achieved by enabling the Serverless feature. For operations and instructions, see Enabling or Disabling Serverless for a Provisioned Resource Cluster.
Can the Configuration of a Serverless Cluster Be Adjusted?
Yes, you can modify the configuration of a Serverless cluster through the console. For details, see Adjusting Configuration.
Does a Serverless Cluster Support Multi-AZ Deployment?
Yes. A Serverless cluster can be deployed across multiple AZs if the following conditions are met: The selected region has at least two AZs. The target AZ has sufficient computing resources. For database version MySQL 5.7, the kernel minor version should be at least 2.0.15. For database version MySQL 8.0, the kernel minor version should be at least 3.0.1.
Which Regions and AZs Support Multi-AZ Deployment for Serverless Clusters?
For details, see Multi-AZ Deployment. If your desired region or AZ is not available, you can submit a ticket to apply to deploy Serverless clusters in the region and AZ.
How Are Refunds Handled for Serverless Clusters?
Since Serverless clusters are billed on a pay-as-you-go basis, fees are calculated based on actual resource usage at the time of settlement. If you have not purchased a resource package, refunds are not involved. Ensure that your account remains funded, as overdue payments may result in cluster isolation and deletion, leading to data loss. For details, see Notes on Arrears.
Can Individual Read-Only Instances Be Manually Started or Stopped?
Yes. Regardless of whether automatic start/stop is enabled for the Serverless cluster, individual read-only instances within the cluster can be manually started or stopped.
Was this page helpful?
You can also Contact Sales or Submit a Ticket for help.
Yes
No
Feedback