Application Real-Time Monitoring Service:Configure a metric trend chart in a Prometheus alert notification

Limits

You can add metric trend charts only to the alert cards of a DingTalk group.

Configure a metric trend chart

1. Log on to the ARMS console.

2. In the navigation pane on the left, choose Managed Service for Prometheus > Prometheus Alert Rules.

3. On the Prometheus Alert Rules page:click Create Prometheus Alert Rule in the upper-right corner of the page.

4. On the create or edit page, perform the following operations:

   1. In the Alert Notification section, select Simple Mode and then select the notification policy that uses the DingTalk group as the contact.

   2. Expand Advanced Settings. In the Annotations section, click Create Annotation to add the following two annotations.

      Key	Value	Example
      _aliyun_display_promql	The PromQL query statement that is used to display the metric trend chart. This statement is usually the PromQL statement for the alert rule with the threshold condition removed. For example, you must change cpu_usage > 80 to cpu_usage. (You can query metric data of up to 512 KB.)	_aliyun_display_promql:cpu_usage
      _aliyun_display_name	The name of the metric in the metric trend chart.	_aliyun_display_name:Container CPU utilization

   3. Modify other parameters as needed. For more information about the parameters, see Prometheus alert rules.

   4. Click Save. Then, click Complete.

View the metric trend chart

After the configuration is complete, when an alert is triggered, you can view the metric trend chart on the alert card in the DingTalk client.

For more information about the details on the alert card, see Handle alerts in a notification group.

FAQ

• Why do I need to manually configure the PromQL query statement for the chart instead of using the PromQL query statement from the alert rule?

  The PromQL query statement for an alert rule is different from the PromQL query statement for the chart. The query statement for an alert rule contains threshold conditions and logical operators such as AND and OR. You must remove these parts from the query statement that is used for the chart.

• Why is the chart not displayed after I configure the _aliyun_display_promql annotation?

  The chart is rendered only if the labels of the time series returned by the _aliyun_display_promql query are identical to the labels of the time series returned by the alert rule's PromQL statement. A PromQL statement can return multiple time series. The system selects one of the time series based on its labels to render the trend chart. For example, if the metric returned by the cpu_usage > 80 statement contains the pod_name and container labels, the result returned by the _aliyun_display_promql query must also contain these two labels.