How Do You Add Failover to a Single-AZ RDS for MySQL Instance?

Why the Answer Is Correct

A Multi-AZ DB instance deployment maintains a synchronously replicated standby in a different Availability Zone, and Amazon RDS automatically fails over to that standby if the primary fails, with the DNS endpoint following the new primary. Because replication is synchronous and the same endpoint is reused, the application keeps performing both reads and writes after a short failover interval, which directly satisfies "provide failover" and "minimize application downtime." No application connection string changes are required, so this is the lowest-effort, lowest-downtime option for a Single-AZ MySQL instance holding mixed read/write traffic. That is why option A is the only solution that both provides failover and keeps writes flowing. The community also converged on this point, as Bad_Mat's vote for option A reflects the standard RDS failover doctrine.

Why the Other Options Are Wrong

A read replica (B) is asynchronous, is read-only, and must be promoted manually; placing it in the same Availability Zone as the primary also means a zonal failure takes both down, so it provides neither automatic failover nor downtime reduction for writes. An Auto Scaling group (C) is an EC2 construct and cannot contain an RDS DB instance at all, so it does nothing for database failover. RDS Proxy (D) pools and multiplexes application connections and can make failover gentler by holding connections open, but it does not itself create a standby or trigger failover — it is a complement to Multi-AZ, not a replacement, and on its own leaves a Single-AZ instance as a single point of failure.

Community Comment Notes

Engagement on this item is light, with Bad_Mat casting the only recorded vote for option A, matching the Multi-AZ DB instance answer. The lack of dissent is typical for this scenario, because the distractors are easy to dismiss once you recall that read replicas and RDS Proxy do not perform failover. Treat the agreement as a sanity check rather than proof, and anchor your reasoning in the synchronous standby behavior of Multi-AZ.