ApsaraMQ for RocketMQ periodically upgrades brokers to improve stability and performance. Brief service interruptions may occur during an upgrade. Review the expected impacts and prepare your applications accordingly.
Before the upgrade
To reduce the impact on production workloads, verify the following before the upgrade begins:
Client retry mechanism: Make sure that automatic retry is enabled on all producer and consumer clients. The built-in retry mechanism handles brief disconnections transparently during the upgrade.
Consumption idempotence: Implement idempotent message processing in your consumer applications. Messages may be delivered more than once during the upgrade. For more information, see Consumption idempotence.
HTTP reconnection logic: If you access ApsaraMQ for RocketMQ over HTTP, make sure that reconnection and retry mechanisms are configured on your clients. HTTP connections may be briefly interrupted or rejected during the upgrade.
Monitoring and alerting: Set up monitoring and alerting for your business metrics before the upgrade begins. For more information, see Monitoring and alerting.
What to expect during the upgrade
Temporary disconnections
The ApsaraMQ for RocketMQ console and service nodes in the cluster may disconnect for several seconds. The disconnection duration is positively correlated with the scale of the cluster. In most cases, the client-side automatic retry mechanism handles these disconnections transparently. However, you may notice:
Exception logs in your application output
Temporary message delivery delays
Duplicate message delivery
Messages may be consumed more than once during the upgrade. Without idempotent message handling, duplicate processing can lead to unintended side effects. See Before the upgrade for preparation steps.
Console feature restrictions
Certain management operations in the ApsaraMQ for RocketMQ console, such as creating or deleting topics, are disabled during the upgrade. Do not perform these operations until the upgrade completes.
HTTP access interruptions
If you access ApsaraMQ for RocketMQ over HTTP, transient connections may occur or connection requests may be rejected. The issue lasts for less than 1 minute. Clients with proper reconnection and retry logic recover automatically.
After the upgrade
After the upgrade completes, verify that:
Producers and consumers have reconnected and are processing messages normally.
No persistent errors appear in your application logs.
Monitoring dashboards show expected throughput and latency.
We apologize for any inconvenience that may be caused by the upgrade. If you experience ongoing issues, submit a ticket to contact support.