All Products
Search

This Product

    Database Autonomy Service:Performance anomaly detection

    Document Center

    Database Autonomy Service:Performance anomaly detection

    Last Updated:Mar 28, 2026

    Database Autonomy Service (DAS) continuously monitors core metrics of your ApsaraDB RDS for MySQL instance using machine learning and intelligent algorithms to detect anomalies and predict issues. When an anomaly appears, use the one-click diagnostics feature to identify the root cause.

    Prerequisites

    Before you begin, make sure that:

    Detect and diagnose performance anomalies

    1. Log on to the DAS console.

    2. In the left-side navigation pane, click Instance Monitoring.

    3. Find the database instance and click the instance ID.

    4. In the left-side navigation pane, click Dashboard, then click the Exception Detection tab.

    5. Specify a time range to view detection and prediction results for that period.

      Note

      The end time must be later than the start time, and the interval cannot exceed seven days.

      • To monitor additional metrics, click More Metrics and select the metrics to include. For a full list of available metrics, see Metrics.

    6. In the Exception Information section, find a metric that shows an anomaly, then click Diagnose in the Diagnose column.

    7. In the Diagnostic Tree dialog box, review the overall performance of the database instance and identify the cause of the exception.

      Exception detection diagnostic tree

    Metrics

    DAS monitors the following metrics for anomaly detection and prediction.

    MetricDescription
    tpsTransactions per second (TPS).
    qpsQueries per second (QPS).
    active_sessionNumber of active sessions.
    delete_psAverage number of DELETE statements executed per second.
    insert_psAverage number of INSERT statements executed per second.
    update_psAverage number of UPDATE statements executed per second.
    select_psAverage number of SELECT statements executed per second.
    bytes_receivedAverage bytes received from all clients per second.
    bytes_sentAverage bytes sent to all clients per second.
    innodb_bp_hitRead hit ratio of the InnoDB buffer pool.
    innodb_data_writtenAverage bytes written to the InnoDB table per second.
    innodb_data_readAverage bytes read from the InnoDB table per second.
    mysql.innodb_log_writesAverage number of physical writes to the InnoDB redo log file per second.
    innodb_rows_deletedAverage rows deleted from the InnoDB table per second.
    innodb_rows_readAverage rows read from the InnoDB table per second.
    innodb_rows_insertedAverage rows inserted into the InnoDB table per second.
    innodb_rows_updatedAverage rows updated in the InnoDB table per second.
    mysql.mem_usageMemory usage of the ApsaraDB RDS for MySQL instance in the entire operating system.
    mysql.cpu_usageCPU utilization of MySQL processes. The maximum value is 100% for Alibaba Cloud database instances.