Global Distributed Cache for Redis features high performance and low latency. You can check the synchronization latency between child instances to obtain the status of synchronization channels and perform O&M operations based on the synchronization latency.

Prerequisites

A distributed instance is created and contains at least two child instances. For more information about how to add child instances, see Add a child instance to a distributed instance.

Procedure

  1. Log on to the ApsaraDB for Redis console.
  2. In the left-side navigation pane, click Global Distributed Cache.
  3. Find the distributed instance that you want to manage and click the Plus icon icon before the instance ID.
    Show the list of child instances
  4. Find the source instance for which you want to query the synchronization latency and click View Synchronization Status in the Actions column.
    Figure 1. View synchronization status
    View synchronization status
  5. In the dialog box that appears, select the time range to query, the data node of the source instance, and the destination instance.
    Note The maximum time range to query is the past day.
    Figure 2. Synchronization status
    Synchronization status
    Metric Description
    current_binlog_sync_delay_time The synchronization latency of binlogs. Unit: seconds. This synchronization latency is caused by the network latency between regions. Generally, a synchronization latency lower than 10s is normal.
    Note If a high-availability switchover is triggered on a child instance, the child instance is restarted, or its configurations are changed, the synchronization latency for this instance may increase to about 60 seconds in a short time. Then, the synchronization latency returns to normal.
    ops The number of binlog entries that are synchronized from the specified data node of the source instance to the destination instance per second.
    status The synchronization status. Valid values:
    • 1: normal
    • 0: abnormal

FAQ

Q: Why does the synchronization latency of binlogs suddenly increase or why is the synchronization status abnormal? How do I troubleshoot the issue?

A: If a high-availability switchover is triggered on a child instance, the child instance is restarted, or its configurations are changed, the synchronization latency for this instance increases, which is a normal fluctuation. The synchronization latency returns to normal after a short time. If the synchronization latency remains high for a long time or the synchronization status is abnormal, you can submit a ticket for help.