An Amazon Aurora cluster uses a single DB instance and the application uses the instance endpoint for both reads and writes. An update is scheduled during a maintenance window and the cluster must remain available with minimal interruption. What should a DevOps engineer do to meet this requirement?

Adding a reader instance to the Aurora cluster and configuring the application to use the cluster (writer) endpoint for writes and the reader endpoint for reads is the correct approach. During a maintenance window, if the primary instance undergoes an update or fails over, the reader instance can be promoted to the new primary, ensuring high availability and minimal downtime. The application, already configured for read/write splitting, will seamlessly continue operations. The other options are incorrect because: Creating a custom ANY endpoint for both reads and writes would still direct all traffic to a single instance, defeating the purpose of high availability during maintenance. Enabling Multi-AZ on an Aurora cluster is not a direct configuration option like with other RDS databases; Aurora achieves high availability through its architecture and reader instances. Even if it were, simply enabling Multi-AZ without read/write splitting wouldn't fully leverage the benefits for minimizing interruption during maintenance.