A company runs an Amazon Aurora DB cluster with a Multi-AZ deployment. The application is configured with a hard-coded endpoint that points to the current writer *DB instance* (an instance-specific endpoint), rather than the Aurora cluster writer endpoint. During an unexpected AZ failure, Aurora promotes the standby to become the new writer. However, the application continues to fail to connect until an operator updates the hard-coded endpoint. What change most directly improves resiliency so the application automatically reconnects after failover?

Keep using the writer DB instance endpoint, but increase the client connection timeout.

Increasing timeouts may reduce transient connection failures, but it does not change which database endpoint the application targets. After failover, the instance-specific endpoint no longer represents the active writer, so connections will still fail.

Disable Multi-AZ failover and rely on manual snapshot restore to bring the database back online.

Disabling failover removes automatic recovery behavior. Manual restore introduces longer downtime and does not provide the immediate endpoint continuity the question is asking for.

Enable cross-Region read replicas and route application traffic to the replica during the outage.

Cross-Region replicas support disaster recovery, but they do not solve automatic AZ failover for the primary writer endpoint within the current Region. Additionally, routing reads/writes to replicas changes the application’s traffic and data-consistency behavior and still requires explicit failover logic at the application or routing layer.

What does this SAA-C03 question test?

Read the scenario before looking for a memorised answer.

What is the correct answer to this question?

The correct answer is: Connect using the Aurora cluster writer endpoint so DNS resolves to the current writer after failover. — Use the Aurora cluster writer endpoint (not an instance-specific endpoint). When Aurora performs a Multi-AZ failover, it promotes a standby to become the writer and updates the cluster writer endpoint so DNS resolution points to the current writer. As a result, the application can reconnect automatically after failover without operators modifying the connection string. Client timeout tuning or changing replica/restore strategies do not provide deterministic endpoint continuity for Multi-AZ failover. Option A delays symptoms but keeps targeting the wrong endpoint after failover. Option C removes automatic failover and increases recovery time because it requires manual intervention. Option D addresses cross-Region disaster recovery, not automatic AZ failover of the Aurora cluster writer endpoint, and would require additional routing and consistency considerations rather than providing seamless reconnection to the writer.

What should I do if I get this SAA-C03 question wrong?

Then try more questions from the same exam bank and focus on understanding why the wrong options are tempting.