Multi-AZ Failover Process
What You'll Learn
You'll discover how Multi-AZ databases automatically recover from failures by switching to a standby instance in another availability zone.
How Automatic Failover Works
When your primary database instance in one availability zone fails, Multi-AZ deployments automatically switch operations to the standby instance in a different zone—all without you lifting a finger.
Automatic Failover Triggers
The system watches for several failure conditions that initiate failover:
- Primary instance hardware failure – Server crashes or becomes unresponsive
- Availability zone outage – The entire data center zone loses connectivity
- Network connectivity loss – Primary can't communicate with the monitoring system
- Storage failure – Disk or volume issues preventing database operations
- Maintenance operations – Planned updates that require instance restart
Think of this like having a backup driver ready to take the wheel the instant your main driver becomes unable to continue—no discussion needed, they just swap seats.
DNS Updates During Failover
The magic happens through DNS. Your application connects using a single endpoint name (like mydb.region.rds.amazonaws.com). During failover, this DNS record automatically updates to point to the standby's IP address instead of the primary's. Your application doesn't need new connection strings—it just reconnects using the same endpoint.
Recovery Time Objectives (RTO)
Typical Multi-AZ failover completes in 1-2 minutes under normal conditions. This includes detection time, promotion of the standby, and DNS propagation. Your application experiences a brief connection interruption, then automatically reconnects to the now-promoted standby.
Key Takeaway: Multi-AZ failover is automatic, triggered by various failure conditions, uses DNS updates to redirect traffic seamlessly, and typically recovers your database within 1-2 minutes with no manual intervention required.