This topic describes how to use the performance insight feature. PolarDB-O provides the diagnosis feature that integrates certain functions of Database Autonomy Service (DAS) and allows you to use the performance insight feature. You can use the performance insight feature to evaluate database loads and identify the root causes of performance issues. This helps you improve database stability.

Background information

The performance insight feature collects and analyzes data from the following sources:
  • If the performance_schema database is used for the target instance, the feature collects and analyzes the data stored in the performance_schema database.
  • If the performance_schema database is not used for the target instance, the feature collects and analyzes the active session data.

Procedure

  1. Log on to the PolarDB console.
  2. On the top of the page, select the region where the target cluster is located.
  3. On the Clusters page, click the ID of the cluster that you want to manage.
  4. In the left-side navigation pane, choose Diagnostics and Optimization > Diagnosis.
  5. On the page that appears, click the Performance Insight tab.
  6. On the Performance Insight tab, click Enable Performance Insight.
    Performance insight
  7. In the message that appears, click Confirm.
  8. On the Performance Insight tab, view and manage the performance information.
    • In the Performance Trend section, you can specify a time range to view the performance metrics of databases. If you want to view the details of a specific performance metric such as CPU usage, click Details next to the performance metric name.Performance trends
      Note The duration of the specified time range cannot exceed seven days.
    • In the Average Active Session section, you can select a session type to view the corresponding trend charts. For example, you can select SQL as the session type. In this section, you can also view the multidimensional details of service loads for each session type. This helps you identify the root causes of performance issues.Active sessionsMultidimensional details of service loads