Configuring Alarms
Last updated:2025-12-24 15:00:43
Scenarios
TDMQ for Apache Pulsar provides multiple monitoring metrics for the running resources to monitor the running status of clusters. It also offers the alarm configuration feature for key metrics. You can configure alarm rules for monitoring metrics. Based on the created alarm rules, the system compares the monitoring metrics against the specified thresholds in a certain period. If a monitoring metric reaches the preset alarm threshold, Tencent Cloud Observability Platform (TCOP) will notify you through emails, Short Message Service (SMS), WeChat, or phone calls. This allows you to take preventive or remedial actions in a timely manner. Proper configuration of alarm rules can help you enhance application robustness and reliability.
Alarm Configuration Recommendations
This section describes key metrics that require special attention when you use TDMQ for Apache Pulsar. It also provides alarm configuration recommendations. You can configure alarm rules based on your business requirements.
Dimension | Alarm Item | Alarm Configuration Recommendation | Impact of Exceeding the Threshold |
Topic | storage_backlog_percentage | 80% | If this threshold is exceeded, messages will fail to be sent, and the client will repeatedly perform reconnections. |
| rate_in_percentage | 80% | By default, a single partition supports 5,000 TPS. If this threshold is exceeded, the server will respond with a latency, and the time for the client to send messages will increase. |
| quota used by the message production traffic | 80% | By default, a single partition supports 40 Mbps. If this threshold is exceeded, the server will respond with a latency, and the time for the client to send messages will increase. |
| Percentage of quota used by consumed messages | 80% | By default, a single partition supports 5,000 TPS. If this threshold is exceeded, a message backlog may occur due to the rate limit of message pushes. |
| throughtput_out_percentage | 80% | By default, a single partition supports 40 Mbps. If this threshold is exceeded, a message backlog may occur due to the rate limit of message pushes. |
| producers_count | 800 | By default, a maximum of 1,000 producers are allowed in a single partition. If this threshold is exceeded, no producers can be created. |
| consumers_count | 1500 | By default, a maximum of 2,000 consumers are allowed in a single partition. If this threshold is exceeded, no consumers can be created. |
Subscription | sub_msg_unack_limit | 80% | By default, a maximum of 5,000 unacknowledged messages are allowed in a single partition for a single subscription. If this threshold is exceeded, the server will stop pushing messages until the client acknowledges all unacknowledged messages. |
| sub_consumer_count | 800 | By default, a maximum of 1,000 consumers are allowed in a single partition for a single subscription. If this threshold is exceeded, no consumers can be created. |
Cluster | [Pro cluster] pulsar_disk_usage | 80% | If this threshold is exceeded, messages will fail to be sent, and writing to the cluster will fail. |
| [Pro cluster] pulsar_tenant_tps_usage | 80% | If this threshold is exceeded, the cluster will run with high load. This poses stability risks. |
| [Pro cluster] tenant_in_limit_count | 1 | When traffic throttling occurs, the server will respond with a latency, resulting in increased sending time. |
Managing Default Alarm Policies
TDMQ for Apache Pulsar provides preset multi-dimensional default alarm policies. You can view and modify these default policies on TCOP. You can also set custom policies based on your business requirements to quickly enable cluster alarm monitoring.
1. Log in to TCOP.
2. In the left sidebar, choose Alarm Management > Alarm Configuration. In the Policy Type section on the left side of the Alarm Policy page, select TDMQ-Pulsar to view the default policies and custom policies under the instance.
3. Select Alarm Type on the left side of the page to view the preset policies and custom policies under the category.
To ensure that the exceptions of your instances can be detected in a timely manner, the default policies are enabled by default when you create an instance. You can click Edit in the operation column corresponding to an alarm policy to view or modify the alarm policy. You can also click Enable or Disable to change the effective status of a policy. For details, see Introduction to Alarm Policies.
4. You can also click Create Policy in the upper-left corner to create a custom policy. For detailed operations and descriptions, see Creating an Alarm Policy.
Configuring Custom Alarm Policies
Step 1: Going to the Configure Alarm Page
TDMQ for Apache Pulsar provides two different entries for configuring alarm rules: configuring alarms on the Monitoring page and configuring alarms on TCOP.
Entry 1: Log in to the TDMQ for Apache Pulsar console. On the Cluster page, click the ID of the target cluster. On the details page, click the Monitoring Information tab, and click the alarm configuration icon in the upper-right corner to redirect to the Configure Alarm page. The default alarm object on the page is the current cluster.
Entry 2: Log in to the TCOP console. On the Alarm Configuration page, you can view all preset policies and custom policies. Click Create Policy, select Cloud Product Monitoring for Monitoring Type, and then select TDMQ/Pulsar for Policy Type.
Step 2: Configuring Alarm Rules
1. After you go to the Alarm Policy Configuration page, select the object for which you want to set alarms.
Tag: You can filter resources for which you want to configure alarm policies in batches based on tags.
Alarm object: Select the TDMQ for Apache Pulsar resources for which alarm policies need to be configured.
2. Set alarm trigger conditions. Select Template and Configure Manually are supported. By default, Configure Manually is selected.
Metric: Take "peak cluster production bandwidth" as an example. If you select the 1-minute statistical granularity, an alarm will be triggered when the peak cluster production bandwidth exceeds the threshold for N consecutive data points within 1 minute.
Alarm Frequency: For example, "alarm triggered once every 30 minutes" indicates that an alarm is triggered once every 30 minutes if a metric exceeds the threshold in multiple consecutive statistical periods. Another alarm will be triggered only if the metric exceeds the threshold again in the next 30 minutes.
1. Select Select Template. Then, click New Trigger Condition Template to redirect to the Trigger Condition Template Setup page.
2. In the upper-left corner, click New Trigger Condition Template. On the New Template page, configure the alarm policy.
Policy Type: Select TDMQ/Pulsar.
Trigger Condition: Set the alarm policy according to the alarm policy recommendations or your actual business requirements.
3. After confirmation, click Save.
4. Return to the Create Alarm Policy page, click Refresh. The alarm policy template that you configured will be displayed.
Note:
3. Click Next step: Configure Alarm Notification to configure alarm recipients.
Step 3: Configuring an Alarm Notification
On the Configure Alarm Notification page, you can select a system preset notification template. In normal cases, the alarm recipient of a preset template is the owner of the root account. If you need to notify the instance owner or other personnel, you can also click Add Notification Template to create a notification template and set alarm recipients and alarm receiving channels.
For detailed operations about how to create a notification template, see Creating a Notification Template.
Was this page helpful?
You can also Contact Sales or Submit a Ticket for help.
Yes
No
Feedback