2.4 Multi-AZ Relational Databases & Multi-Region DR Strategies

🎯 Key AWS SAA-C03 Architectural Takeaways

• Amazon RDS Multi-AZ synchronously replicates to a standby instance in another AZ, providing automated 60-120 second DNS failover with zero data loss (RPO = 0).
• RDS Read Replicas replicate asynchronously and are designed for read scaling; promoting a read replica during an outage incurs data loss (RPO > 0).
• Amazon Aurora Global Database replicates storage across regions with typical lag under 1 second and enables sub-minute cross-region disaster recovery promotion.
• The four AWS disaster recovery tiers in order of increasing cost and decreasing RTO/RPO are Backup & Restore, Pilot Light, Warm Standby, and Multi-Site Active-Active.

High availability in the database layer begins with Amazon RDS Multi-AZ deployments. When Multi-AZ is enabled, RDS provisions a primary database instance in one AZ and a synchronous standby replica in a second AZ. Every database write is synchronously committed to both storage volumes before acknowledging the client. If an AZ outage occurs, RDS automatically changes the database DNS CNAME to the standby within 60-120 seconds with zero data loss.

For globally distributed relational workloads, Amazon Aurora Global Database uses dedicated storage-level infrastructure to replicate data across up to five secondary AWS Regions. Typical cross-region replication lag is under 1 second. In the event of a regional catastrophe, a secondary Aurora cluster can be promoted to full read/write production in under 1 minute.

At the DNS layer, Amazon Route 53 Failover routing implements active-passive disaster recovery. An Alias record for the primary application is associated with a Route 53 health check. If the health check detects that the primary endpoint is down, Route 53 automatically diverts all DNS query responses to the secondary disaster recovery endpoint (such as an S3 maintenance page or secondary region ALB).

Choosing the appropriate disaster recovery strategy depends on business Recovery Time Objectives (RTO) and Recovery Point Objectives (RPO). While Backup and Restore takes hours at minimal cost, Pilot Light keeps core database replication running live while maintaining compute instances dormant, achieving an RTO under tens of minutes at a fraction of the cost of running an Active-Active duplicate fleet.

⚠️ Common Pearson VUE / AWS Exam Traps

• Confusing RDS Multi-AZ standby (synchronous, failover only, cannot serve read traffic) with Read Replicas (asynchronous, read-only traffic, manual promotion).
• Assuming Route 53 Weighted routing provides active-passive disaster recovery — weighted routing sends live traffic continuously to both endpoints; Failover routing is required for active-passive.
• Over-engineering a system with Multi-Site Active-Active when the business requirements tolerate an RTO of 30 minutes (where Pilot Light is vastly cheaper).