Configuring a Recording Rule to Accelerate Metric Query

Scenario

Some metrics may require much calculation on the query end, affecting query performance. You can configure recording rules to calculate common or complex metrics in advance. Especially in large-scale clusters and complex service scenarios, recording rules can reduce PromQL complexity, improve metric query performance, and prevent slow configuration and query.

Practice

Configure a recording rule to perform secondary development on metrics. Then view recording rule metrics on the Metric Browsing page.

Configuring a Recording Rule

1. Log in to the AOM 2.0 console.
2. In the navigation pane on the left, choose Prometheus Monitoring > Instances.
3. Click Add Prometheus Instance. For details about the parameters, see Table 1.

   Table 1 Parameters for adding a common Prometheus instance

   Parameter	Description	Example Value
   Instance Name	Prometheus instance name. Enter a maximum of 100 characters and do not start or end with an underscore (_) or hyphen (-). Only letters, digits, underscores, and hyphens are allowed.	mon_AOM
   Enterprise Project	Enterprise project. If Enterprise Project is set to All on the global settings page, select an enterprise project from the drop-down list here. If you have already selected an enterprise project on the global settings page, this option will be grayed and cannot be changed.	default
   Instance Type	Type of a Prometheus instance. Only Prometheus instances for CCE and common Prometheus instances support recording rules.	Common Prometheus instance

4. Click OK.
5. In the Prometheus instance list, click the common Prometheus instance created in 3.
6. In the navigation pane on the left, choose Settings. In the Recording Rules area, click Edit RecordingRule.yaml.
7. In the dialog box that is displayed, delete the default content and enter a custom recording rule.
   Only one RecordingRule.yaml file needs to be configured for a cluster. Each rule group name must be unique.

   Figure 1 Configuring a recording rule
   

   Table 2 Recording rule parameters

   Parameter	Description
   groups	Rule group. You can set multiple rule groups in one RecordingRule.yaml file.
   name	Rule group name. Each rule group name must be unique.
   interval	Execution interval of a rule group. The default value is 60s. (Optional. It is recommended that the execution interval be the same as the metric reporting period.)
   rules	Rule. A rule group can contain multiple rules.
   record	Rule name. It must comply with Prometheus metric name specifications.
   expr	Calculation expression, which is used to calculate metric values. It must comply with PromQL requirements.
   labels	(Optional) Metric label. Labels must comply with Prometheus metric label specifications.

   Example of a recording rule:

   groups:
     - name: apiserver_request_total
       interval: 60s
       rules:
         - record: apiserver_request_rate
           expr: avg by (job, instance, mode) (rate(apiserver_request_total[5m]))
           labels:
             team: operations
         - record: job:apiserver_request_total:sum_rate10m
           expr: sum by (job)(rate(apiserver_request_total[10m]))
           labels:
             team: operations

8. Click OK.

Viewing Recording Rule Metrics

After the recording rule is configured, you can view recording rule metrics on the Metric Browsing page of AOM or through Grafana. The following shows how to view recording rule metrics on the Metric Browsing page of AOM.

1. Log in to the AOM 2.0 console.
2. In the navigation pane, choose Metric Browsing.
3. On the Metric Sources tab page, select the common Prometheus instance created in 3.
4. Choose All Metrics and Multiple Metrics, and enter a recording rule metric name in the search box to view details.