ApsaraMQ for RocketMQ 5.x provides two tiers of disaster recovery: multi-AZ (availability zone) high availability and cross-region disaster recovery. Multi-AZ high availability protects against single-zone failures. Cross-region disaster recovery protects against regional outages by replicating messages and consumer progress to a different region.
Note: Multi-AZ high availability handles failures in a single availability zone. To protect against regional disasters, see Cross-region high availability.
Multi-AZ high availability
Multi-AZ high availability distributes services and data across multiple availability zones within the same region. If one availability zone fails, the remaining zones automatically handle the workload without interrupting message processing.
This tier has two layers: service high availability and data high availability.
Service high availability
All ApsaraMQ for RocketMQ instances except legacy single-node instances provide multi-AZ service high availability by default—no configuration required.
If a service disruption occurs in one availability zone, the client automatically redirects message requests to a healthy zone. For example, if AZ1 becomes unavailable, clients connected through the console endpoint fail over to AZ2 or AZ3, and producers and consumers continue operating without manual intervention.
Data high availability
Dedicated serverless instances and Platinum Edition instances (both subscription and pay-as-you-go) replicate data synchronously across availability zones.
The mechanism uses zone-redundant cloud disks combined with ECS multi-disk attachment to replicate writes in real time. If one zone fails, consumers in the surviving zones can process any accumulated messages.
Disaster recovery drill
Use the fault drill for availability zones described in Fault drills to simulate AZ-level failures before they occur in production. Running drills validates your high-availability design and failover effectiveness.
Cross-region high availability
Use Global Replicator to configure synchronization tasks that replicate messages and consumer progress between instances in different regions. Synchronizing consumer progress ensures that when consumers switch to a different region during a failover, they resume from the correct position.
For more information, see Scenarios.