- Uses synchronous replication between a primary and standby DB in different AZs.
- Single DNS endpoint — automatic failover to standby on failure.
- Improves availability and resilience against:
- AZ failures
- Network issues
- Instance or storage failures
- Failover is automatic, no application changes required.
- Not used for read scaling (standby cannot serve reads).
- Read Replicas can be configured as Multi-AZ for combined DR and scalability.
Architecture
- Application connects to one DNS name.
- Primary handles all reads/writes.
- Standby in another AZ is kept in sync via synchronous replication.
- On failure, DNS switches to the standby automatically.
Upgrading from Single-AZ to Multi-AZ
- Zero downtime process.
- Steps performed by AWS:
- Take a snapshot of the existing DB.
- Restore snapshot in a different AZ to create standby.
- Establish synchronous replication between primary and standby.