All Products
Search

This Product

    PolarDB:Performance trend

    Document Center

    PolarDB:Performance trend

    Last Updated:Jul 21, 2023

    This topic describes how to use the performance trend feature to view the performance metrics of your PolarDB-X instance.

    PolarDB-X is integrated with the performance trend feature that is provided by Data Autonomy Service (DAS). You can use this feature to view the details of each performance metric of compute nodes, data nodes, Global Meta Service (GMS) nodes, and change data capture (CDC) nodes of your PolarDB-X instance. You can use multiple methods to view the performance trends of your PolarDB-X instance. You can specify a time range to view the performance metrics that are collected in the specified period of time. You can specify two dates to view the comparison between the performance metrics that are collected on the specified days. You can also use custom dashboards to display the performance metrics that you want to view.

    View performance trends

    1. Log on to the PolarDB for Xscale console.
    2. In the left-side navigation pane, click Instance List. On the Instance List page, click the PolarDB-X 2.0 tab.
    3. In the top navigation bar, select the region in which your instance resides.
    4. Find the instance that you want to manage and click the instance ID.
    5. In the left-side navigation pane, choose Diagnosis and optimization > Performance trend.
      • On the page that appears, you can view the performance metrics of the compute nodes, data nodes, GMS nodes, and CDC nodes. For more information about each metric, see Metric description.

        In the Nodes section of each tab, you can view information about the nodes, including the specifications, CPU utilization, memory usage, and current active connections.

      • Click the Performance Trends tab, specify a node and a time range, and then click Search to view performance monitoring data in each chart.
        • You can click More Metrics to select the metrics whose performance trend that you want to view.
        • In a performance trend chart, drag and drop the pointer to specify a time range and click Diagnose to check whether the metric is normal within the selected time range. For more information, see Diagnose node performance.
        • You can click Details in the upper-right corner of the trend chart of a metric to zoom in the trend chart. In this case, you can modify the time to view the trend of the metric over time.
      • Click the Performance Trend Comparison tab. Specify a node, two dates, and a time range, and then click Search to view the comparison between the performance metrics that are collected on the specified days.
      • On the Custom Chart tab, you can create dashboards to display the performance metrics that you want to view. When you create a chart, you can specify multiple performance metrics that you want to display on the chart. This way, you can analyze metrics that are displayed on the dashboards and troubleshoot performance issues.
        Note You must create dashboards before you create charts. For more information, see Monitoring dashboards.

    Diagnose node performance

    PolarDB-X provides the performance diagnostics feature. In a performance trend chart of a node, select a time range to diagnose the performance of the node based on resource usage and slow SQL queries.

    Note Compute nodes and data nodes support the performance diagnostics feature. GMS nodes and CDC nodes do not support this feature.
    1. Log on to the PolarDB for Xscale console.
    2. In the left-side navigation pane, click Instance List. On the Instance List page, click the PolarDB-X 2.0 tab.
    3. In the top navigation bar, select the region in which your instance resides.
    4. Find the instance that you want to manage and click the instance ID.
    5. In the left-side navigation pane, choose Diagnosis and optimization > Performance trend.
    6. In a performance trend chart, drag and drop the pointer to specify a time range and click Diagnose.
    7. On the Details page that appears, view the statistical information about resource usage and slow SQL queries.

    Metric description

    Compute nodes
    MetricUnitDescription
    polardbx.cpu_usage%The average CPU utilization within the specified period of time.
    polardbx.mem_usage%The memory usage of the Java Virtual Machine (JVM). In specific scenarios, the memory usage fluctuates.
    polardbx.active_connectionCountThe total number of connections.
    polardbx.running_threadCountThe number of active connections.
    polardbx.network_in_bytesbyteThe total amount of inbound traffic.
    polardbx.network_out_bytesbyteThe total amount of outbound traffic.
    polardbx.logic_qpsPer SecondThe total number of logical SQL query statements that are executed per second. Logical SQL query statements are SQL statements that are sent from your business application to your PolarDB-X database.
    polardbx.physical_qpsPer SecondThe total number of physical SQL query statements that are executed per second. Physical SQL query statements are executed on the read-only nodes of your database.
    polardbx.logic_rtMillisecondsThe average response time of logical SQL queries.
    polardbx.physical_rtMillisecondsThe average response time of physical SQL queries.
    polardbx.slow_request_countPer SecondThe number of slow logical SQL queries.
    polardbx.physical_slow_request_countPer SecondThe number of slow physical SQL queries.
    Data nodes and GMS nodes
    MetricUnitDescription
    mysql.tpsPer SecondThe number of transactions that are run on the data node per second.
    mysql.qpsPer SecondThe number of queries that are run on the data node per second.
    mysql.total_sessionCountThe total number of sessions.
    mysql.active_sessionCountThe number of active sessions.
    mysql.bytes_receivedKBThe average number of bytes in the data that the node receives from all clients per second.
    mysql.bytes_sentKBThe average number of bytes in the data that is sent from the node to all clients per second.
    mysql.tb.tmp.diskCountThe number of temporary tables that are automatically created on the disk of the node when SQL statements are executed.
    mysql.insert_psPer SecondThe average number of INSERT statements that are executed per second.
    mysql.select_psPer SecondThe average number of SELECT statements that are executed per second.
    mysql.update_psPer SecondThe average number of UPDATE statements that are executed per second.
    mysql.delete_psPer SecondThe average number of DELETE statements that are executed per second.
    mysql.replace_psPer SecondThe average number of REPLACE statements that are executed per second.
    mysql.innodb_data_writtenKBThe average number of bytes in the data that is written to InnoDB per second.
    mysql.innodb_data_readKBThe average number of bytes in the data that is read from InnoDB per second.
    mysql.innodb_buffer_pool_reads_requestsCountThe average number of logical reads that read data pages from the InnoDB buffer pool per second.
    mysql.innodb_bp_dirty_pct%The dirty page ratio of the InnoDB buffer pool. Formula: Innodb_buffer_pool_pages_dirty/Innodb_buffer_pool_pages_data × 100%
    mysql.innodb_bp_hit%The read 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_writesPer SecondThe average number of physical writes to the InnoDB redo log file per second.
    mysql.innodb_os_log_fsyncsPer SecondThe average number of fsync() writes to the log file per second.
    mysql.innodb_rows_deletedPer SecondThe average number of InnoDB rows from which data is deleted per second.
    mysql.innodb_rows_readPer SecondThe average number of InnoDB rows from which data is read per second.
    mysql.innodb_rows_insertedPer SecondThe average number of InnoDB rows into which data is inserted per second.
    mysql.innodb_rows_updatedPer SecondThe average number of InnoDB rows in which data is updated per second.
    mysql.mem_usage%The memory usage of the MySQL instance in the entire operating system.
    mysql.cpu_usage%The CPU utilization of MySQL processes. This metric can reach up to 100% for Alibaba Cloud database instances.
    mysql.data.sizeMBThe size of the data space.
    mysql.tmp.sizeMBThe size of the temporary tablespace.
    mysql.other.sizeMBThe system space.
    mysql.instance.sizeMBThe total used space of the MySQL database instance.
    mysql.log.sizeMBThe size of the log space.
    mysql.iopsCountThe IOPS of the MySQL instance.
    CDC nodes
    MetricUnitDescription
    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_delayMillisecondsThe latency for processing binary log events.