Description
By default, this alert is triggered if the CPU utilization of an OBServer node remains higher than 90% for more than 5 minutes and the CPU utilization of an SQL statement on the OBServer node exceeds 15%.
Principle
| Parameter | Value |
|---|---|
| Metric | None |
| Source | None |
| Collected metric | None |
| Metric expression | None |
| Collection cycle | None |
Alert rule
By default, this alert is triggered if the CPU utilization of an OBServer node remains higher than 90% for more than 5 minutes and the CPU utilization of an SQL statement on the OBServer node exceeds 15%.
Alert information
| Trigger method | Alert level | Scope |
|---|---|---|
| Based on monitored conditions | Critical | Cluster |
Alert templates
Details
Template:
The CPU utilization of the host ${svr_ip} in the ${ob_cluster_name} cluster reaches ${server_cpu_percent}% during the period from ${start_time} to ${end_time}. The CPU utilization of the SQL statement ${sql_id} in the ${db_name} database of the ${tenant_name} tenant is too high. The CPU utilization is ${sql_cpu_percent}%. The average CPU time is ${sql_cpu_time_ms} ms. The average response time is ${sql_elapsed_time_ms} ms. Recommended action: [Throttle].Example:
Overview: alarm_template_id=0:ob_cluster=admin-4:tenant_name=a****:db_name=test:sql_id=142A49D034B27871DE8E646F3360****:recommend_operations=RATE_LIMIT. The CPU utilization of the host exceeds the threshold because the CPU utilization of the SQL statement is too high. Period: 2023-09-06T00:03:17 to 2023-09-06T00:08:47 Symptom: The CPU utilization of the OBServer node xxx.xxx.xxx.xxx reaches 22.08%. Associated SQL statement: SQL ID: 142A49D034B27871DE8E646F3360****. Database: test. Tenant: a****. Hit rule: The CPU utilization of the SQL statement is too high. Basic information of the SQL statement: The CPU utilization of the SQL statement is 17.84%. The average CPU time is 101.95 ms. The average response time is 1313.44 ms. Recommended action: [Throttle].
Impact on the system
The CPU utilization of the OBServer node is affected.
Possible causes
The SQL statement is poorly or frequently executed.
Solutions
- Optimize the SQL statement.
- Perform throttling based on the business assessment.