How to view the performance trend of PolarDB-X - PolarDB

PolarDB-X integrates the performance trend feature of Data Autonomy Service (DAS) to provide monitoring metrics for compute nodes (CN), data nodes (DN), GMS nodes, and change data capture (CDC) nodes. You can view performance trends for a specified time range, compare trends between two periods, and create custom charts.

Viewing performance trends

Log on to the PolarDB Distributed Edition console.
On the Instances page, click the PolarDB-X 2.0 tab.
In the upper-left corner of the page, select the region where your target instance is located.
Find your target instance and click its instance ID.
In the left-side navigation pane, choose Diagnostics and Optimization > Performance Trend.
- Click the Compute Node (CN), Data Node (DN), GMS node, and Change Data Capture (CDC) Node tabs to view their monitoring metrics. For more information, see Performance metrics.
  Note
  - A PolarDB-X Standard Edition instance supports only Data Node (DN).
  - The node list displays information such as node specifications, CPU utilization, memory usage, and the current number of active connections.
- On the Performance Trend tab, view performance trends for various metrics for the selected time range.
  - Click More Metrics to select additional metrics whose performance trends you want to view.
  - In any metric trend chart, drag the cursor over a time range to perform a performance diagnosis and check the metric status for that period.
  - Click Details in a metric's trend chart to zoom in on it. You can also change the time range to view the metric's trend over different periods.
- On the Performance Trend Comparison tab, set two time ranges and click View to view a comparative chart of performance trends.
- On the Custom Charts tab, create custom monitoring dashboards to display multiple performance metrics in a single chart, simplifying troubleshooting and analysis.
  
  Note
  If this is your first time using this feature, you must create a monitoring dashboard first.

Performance diagnosis

In addition to viewing performance trends, you can diagnose a node's performance for a specific period by dragging the cursor over a time range in its trend chart. The diagnosis is based on its resource utilization and slow SQL queries.

Note

Currently, performance diagnosis is available only for compute nodes and data nodes. This feature is not available for GMS nodes or change data capture (CDC) nodes.

Log on to the PolarDB Distributed Edition console.
On the Instances page, click the PolarDB-X 2.0 tab.
In the upper-left corner of the page, select the region where your target instance is located.
Find your target instance and click its instance ID.
In the left-side navigation pane, choose Diagnostics and Optimization > Performance Trend.
In a node's trend chart, drag the cursor to select a time range, and then click the Diagnose button.
On the diagnosis details page, wait for the task to complete. You can then view any anomalies in resource utilization and slow SQL queries.

Performance metrics

Compute nodes

Metric	Unit	Description
polardbx.cpu_usage	%	The average CPU utilization.
polardbx.mem_usage	%	The memory usage of the Java Virtual Machine (JVM). Fluctuations in memory usage are normal.
polardbx.active_connection	Count	The total number of established connections.
polardbx.running_thread	Count	The number of threads currently executing queries.
polardbx.network_in_bytes	Byte	The total volume of inbound network traffic.
polardbx.network_out_bytes	Byte	The total volume of outbound network traffic.
polardbx.logic_qps	QPS	The total number of logical SQL statements processed per second.
polardbx.physical_qps	QPS	The total number of physical SQL statements processed per second.
polardbx.logic_rt	ms	The average response time for logical SQL queries.
polardbx.physical_rt	ms	The average response time for physical SQL queries.
polardbx.slow_request_count	requests/s	The number of logical slow SQL queries per second.
polardbx.physical_slow_request_count	requests/s	The number of physical slow SQL queries per second.

Data nodes and GMS nodes

Metric	Unit	Description
mysql.tps	TPS	The number of transactions per second (TPS).
mysql.qps	QPS	The number of queries per second (QPS).
mysql.total_session	Count	The total number of current sessions.
mysql.active_session	Count	The number of current active sessions.
mysql.bytes_received	KB/s	The average number of bytes received from all clients per second.
mysql.bytes_sent	KB/s	The average number of bytes sent to all clients per second.
mysql.tb.tmp.disk	Count	The number of temporary tables automatically created on disk when MySQL executes statements.
mysql.insert_ps	statements/s	The average number of INSERT statements executed per second.
mysql.select_ps	statements/s	The average number of SELECT statements executed per second.
mysql.update_ps	statements/s	The average number of UPDATE statements executed per second.
mysql.delete_ps	statements/s	The average number of DELETE statements executed per second.
mysql.replace_ps	statements/s	The average number of REPLACE statements executed per second.
mysql.innodb_data_written	KB/s	The average number of bytes written to InnoDB per second.
mysql.innodb_data_read	KB/s	The average number of bytes read from InnoDB per second.
mysql.innodb_buffer_pool_reads_requests	requests/s	The average number of logical reads from the InnoDB buffer pool per second.
mysql.innodb_bp_dirty_pct	%	The percentage of dirty pages in the InnoDB buffer pool. Formula: Innodb_buffer_pool_pages_dirty / Innodb_buffer_pool_pages_data × 100%.
mysql.innodb_bp_hit	%	The cache hit ratio of the InnoDB buffer pool. Formula: (Innodb_buffer_pool_read_requests - Innodb_buffer_pool_reads) / Innodb_buffer_pool_read_requests × 100%.
mysql.innodb_bp_usage_pct	%	The usage of the InnoDB buffer pool. Formula: innodb_buffer_pool_pages_data / (innodb_buffer_pool_pages_data + innodb_buffer_pool_pages_free) × 100%.
mysql.innodb_log_writes	writes/s	The average number of physical writes to the InnoDB redo log file per second.
mysql.innodb_os_log_fsyncs	fsyncs/s	The average number of fsync() writes to the log file per second.
mysql.innodb_rows_deleted	rows/s	The average number of rows deleted from InnoDB tables per second.
mysql.innodb_rows_read	rows/s	The average number of rows read from InnoDB tables per second.
mysql.innodb_rows_inserted	rows/s	The average number of rows inserted into InnoDB tables per second.
mysql.innodb_rows_updated	rows/s	The average number of rows updated in InnoDB tables per second.
mysql.mem_usage	%	The memory usage of the MySQL instance as a percentage of the total operating system memory.
mysql.cpu_usage	%	The CPU utilization of the MySQL service process. For Alibaba Cloud database instances, this metric can reach a maximum of 100%.
mysql.data.size	MB	The space used by data.
mysql.tmp.size	MB	The space used by temporary tables.
mysql.other.size	MB	The space used by system files.
mysql.instance.size	MB	The total space used by the MySQL instance.
mysql.log.size	MB	The space used by log files.
mysql.iops	IOPS	The number of read and write I/O operations per second.

Change data capture (CDC) nodes

Metric	Unit	Description
polardbx_cdc.cpu_usage	%	The CPU utilization.
polardbx_cdc.mem_usage	%	The memory usage.
polardbx_cdc.dumper_heapUsage	%	The heap memory usage.
polardbx_cdc.dumper_delay	ms	The latency in processing binary log events.