The PolarDB console allows you to monitor a variety of performance metrics and view monitoring data at intervals of seconds. You can monitor the status of your clusters and locate faults based on the monitoring data.

Procedure

  1. Log on to the PolarDB console.
  2. In the upper-left corner of the console, select the region where the cluster that you want to manage is deployed.
  3. Find the cluster you want to manage and click the cluster ID.
  4. In the left-side navigation pane, choose Diagnostics and Optimization > Monitoring.
  5. View the monitoring information about the cluster on the Cluster, PolarProxy, Compute Node, or Advanced Monitoring tab. For more information, see Metric description.
    • To monitor cluster performance, click the Cluster tab, specify a period of time on the right side of the page, and then click OK. Cluster performance monitoring
    • To monitor database proxy performance, click the PolarProxy tab, specify a period of time on the right side of the page, and then click OK. Database proxy
    • To monitor compute node performance, click the Compute Node tab, select a node from the drop-down list on the right side of the page, specify a period of time in the text box next to the drop-down list, and then click OK. Node performance monitoring
Note On the Real-time Monitoring Dashboard page in the Database Autonomy Service (DAS) console, you can view the monitoring data about PolarDB for MySQL. This can help you identify abnormal clusters for further analytics and optimization.

Metrics

Category Metric Description
Cluster Storage Displays the sizes of log files such as binary log and redo log files, and the storage used by data files, system files, and temporary files. Contact Alibaba Cloud technical support to obtain this metric.
QPS Displays the queries per second (QPS) of each node. The values are sourced from the Questions metric in the output of the show global status statement.
TPS Displays the transactions per second (TPS) of each node. The values are sourced from the Com_commit and Com_rollback metrics in the output of the show global status statement.
MPS Displays the manipulations per second (MPS) of each node. The values are sourced from the Com_insert, Com_insert_select, Com_update, Com_update_multi, Com_delete, Com_delete_multi, Com_replace, and Com_replace_select metrics in the output of the show global status statement.
CPU Usage Displays the CPU utilization of each node. Contact Alibaba Cloud technical support to obtain this metric.
Memory Usage Displays the memory usage of each node. Contact Alibaba Cloud technical support to obtain this metric.
Connections Displays the number of active connections per second on each node. The values are sourced from the statistical information in the output of the show processlist statement.
Compute node QPS Displays the QPS of the selected node. The values are sourced from the Questions metric in the output of the show global status statement.
TPS Displays the TPS of the selected node. The values are sourced from the Com_commit and Com_rollback metrics in the output of the show global status statement.
MPS Displays the manipulations per second (MPS) of each node. The values are sourced from the Com_insert, Com_insert_select, Com_update, Com_update_multi, Com_delete, Com_delete_multi, Com_replace, and Com_replace_select metrics in the output of the show global status statement.
CPU Usage Displays the CPU utilization of the selected node. Contact Alibaba Cloud technical support to obtain this metric.
Memory Usage Displays the memory usage of the selected node. Contact Alibaba Cloud technical support to obtain this metric.
Connections Displays the total number of connections and the number of active connections on the selected node. The values are sourced from the statistical information in the output of the show processlist statement.
Operations Displays the number of operations performed per second on the selected node. The operations counted include the DELETE, INSERT, UPDATE, and REPLACE operations. The values are sourced from the metrics whose names start with Com_ in the output of the show global status statement.
Memory Buffer Pool Displays the dirty ratio, read hit ratio, and usage of the buffer pool on the selected node. The values are sourced from Innodb_buffer_pool_pages_dirty, Innodb_buffer_pool_pages_total, Innodb_buffer_pool_reads, Innodb_buffer_pool_read_requests, and other metrics in the output of the show global status statement.
I/O Throughput Displays the total I/O throughput, read I/O throughput, and write I/O throughput of the selected node. Contact Alibaba Cloud technical support to obtain this metric.
IOPS Displays the total IOPS, read IOPS, and write IOPS of the selected node. Contact Alibaba Cloud technical support to obtain this metric.
Network Displays the input traffic per second and output traffic per second of the selected node. The values are sourced from the Bytes_received and Bytes_sent metrics in the output of the show global status statement.
Scanned Rows Displays the numbers of rows that are inserted, read, updated, and deleted per second on the selected node. The values are sourced from the Innodb_rows_deleted, Innodb_rows_inserted, Innodb_rows_read, and Innodb_rows_updated metrics in the output of the show global status statement.
InnoDB Read and Written Data Displays the amount of data that is read from or written into the storage engine per second on the selected node. The values are sourced from the Innodb_data_read and Innodb_data_written metrics in the output of the show global status statement.
InnoDB Buffer Pool Requests Displays the number of read operations and the number of write operations that are performed per second on the buffer pool of the selected node. The values are sourced from the Innodb_buffer_pool_read_requests and Innodb_buffer_pool_write_requests metrics in the output of the show global status statement.
InnoDB Log Writes Displays the number of log write requests per second and the number of times that data is synchronized to disks per second on the selected node. The values are sourced from the Innodb_log_write_requests and Innodb_os_log_fsyncs metrics in the output of the show global status statement.
Temporary Tables Displays the number of temporary tables that are created per second on the selected node. The values are sourced from the Created_tmp_disk_tables metric in the output of the show global status statement.

FAQ

  • Why is the QPS that is displayed on the Monitoring page approximately 10 when the service traffic of my cluster is 0?

    The monitoring, log collection, and administration tasks that run in the system background generate approximately 10 queries per second. This has little impact on the performance of your cluster.

  • How do I reduce the CPU utilization when the CPU utilization is excessively high?
    We recommend that you perform the following steps:
    • Check whether a large number of slow requests exist. If a large number of slow requests exist, we recommend that you optimize slow SQL statements first. For more information about how to view and optimize slow SQL statements, see Slow SQL query.
    • Check whether the trend of the CPU utilization curve is consistent with that of the QPS or TPS curve. If the curves are consistent, the issue is caused by high concurrency of transactions. In this case, we recommend that you upgrade the configuration of your cluster. For more information about how to upgrade the configuration of a cluster, see Manually upgrade or downgrade a PolarDB cluster.
    • If no slow requests exist and the trend of the CPU utilization curve is inconsistent with that of the QPS or TPS curve, Submit a ticket for technical support.
  • What do I do when the number of connections is much greater than the number of active connections?

    You can specify smaller values for the wait_timeout and interactive_timeout parameters to accelerate the release of idle connections. We recommend that you close the connections that are no longer in use in time to reduce the number of idle connections.

Related API operations

API Description
DescribeDBClusterPerformance Queries the performance data of a specified PolarDB cluster.
DescribeDBNodePerformance Queries the performance data of a node in a specified PolarDB cluster.
DescribeDBClusterMonitor Queries the interval at which the monitoring data of a specified PolarDB cluster is collected.