This topic describes the performance monitoring and diagnostics for OceanBase clusters.
Monitor cluster performance
Procedure
Log on to the OCP console. In the left-side navigation pane, click Clusters.
In the Clusters list on the Clusters page, find the target cluster and click its name.
In the left-side navigation pane on the page that appears, click Performance Monitoring.
Enable real-time performance monitoring to view the real-time performance changes.
The following table describes the metric data filter criteria.
| Filter criterion | Description |
|---|---|
| Select Time | The time range of the metric data to be displayed. You can view data for the last 30 days. |
| Stat Period | The statistical period of metric data. You can choose to calculate and display the average values of collected metric data by minute or second. By default, data is displayed by minute. |
| Zone | The zone whose metric data you want to view. OceanBase Database is a distributed database service, and cluster metrics are collected and aggregated from multiple zones. The aggregation method depends on the metric type. You can select a zone to view metric data of the zone. |
| OBServer | The OBServer node whose metric data you want to view. OceanBase Database is a distributed database service, and cluster metrics are collected and aggregated from multiple OBServer nodes. The aggregation method depends on the metric type. You can select an OBServer node to view its metrics. |
Performance metrics are classified into database metrics and host metrics.
Cluster performance report
Procedure
Log on to the OCP console. In the left-side navigation pane, click Clusters.
In the Clusters list on the Clusters page, find the target cluster and click its name.
In the left-side navigation pane of the page that appears, click Performance Report.
Click Generate Performance Report. Specify the snapshot start and end points to generate a cluster performance report.
Click View to view the performance report. The system retains the data for the last 90 days.
Click Download to download the performance report.
The following table describes the content of a performance report.
| Module | Description |
|---|---|
| Basic information | This module shows the version, OBServer nodes, tenants, and data partitions of the cluster. |
| Workloads | This module shows the workloads of the cluster. |
| SQL | This module shows the top SQL information aggregated by metrics such as response time, wait event, and CPU utilization. |
| Deadlock | This module shows the row lock requests, waits, and timeouts. |
View performance data by using views
| Views | Description |
|---|---|
| v$sysstat and gv$sysstat | These views display system statistics by server or cluster, such as queries per second (QPS), transactions per second (TPS), input/output operations per second (IOPS), and the cache hit rate. |
| v$latch and gv$latch | These views display the number of times latches are added, the number of spin operations performed by latches, and the waiting duration of latches. They can help you identify system hotspots. |
| v$lock_wait_stat and gv$lock_wait_stat | These views display information about row locks. They can help you identify lock conflicts and hotspot rows in the system. |
| v$session_wait, gv$session_wait, v$session_wait_history, and gv$session_wait_history | These views display the number and the waiting duration of wait events, such as disk I/O wait events and remote procedure call (RPC) wait events, by session. They can help you identify the causes of slow SQL executions and identify system bottlenecks. |