An application uses an Amazon Aurora DB cluster. The cluster performs an automatic failover from the writer instance to a standby instance. After failover completes, reads succeed, but all new writes fail with errors indicating the application is connecting to the old writer endpoint. Which change best fixes the resiliency issue after failover?

Enable Multi-AZ on the individual writer instance settings so it can automatically create a new instance during failover.

For Aurora, failover behavior and Multi-AZ/replication are managed at the cluster level. Changing an individual instance’s Multi-AZ settings does not correct the fundamental problem that the application is connecting to the wrong endpoint for the current writer role.

Increase the failover timeout for Aurora to 60 minutes to ensure the app finishes reconnecting.

Failover timeout affects how long Aurora may take to complete the failover process, but it does not fix the client-side endpoint selection. If the application continues to target the old writer instance endpoint, writes will keep failing after failover as well.

Switch the cluster to a single-AZ configuration to reduce connection retries after failover.

A single-AZ configuration reduces availability and removes the intended Aurora high-availability behavior. It also does not resolve the endpoint mismatch causing writes to fail.

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: Update the application to use the Aurora cluster writer endpoint (or the cluster endpoint intended for writes) rather than an instance-specific endpoint. — Aurora failover changes which underlying DB instance is the writer. Resilient clients should not hard-code instance-specific hostnames. Using the Aurora cluster writer endpoint (cluster endpoint intended for writes) ensures the application connects to whichever instance currently has the writer role after failover. If the application instead uses an instance-specific endpoint from before failover, writes fail until the configuration is corrected. B) Aurora’s Multi-AZ/high availability is cluster-managed; instance-level changes do not address stale endpoint usage. C) Increasing timeouts only changes when failures happen, not which endpoint the application uses. D) Single-AZ undermines resiliency and still does not fix the client using the old writer endpoint.

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.