An Auto Scaling group of On-Demand EC2 instances processes messages from an SQS queue. Each message can take up to 12 hours to finish processing. The Auto Scaling group scales based on the queue length. How should the SysOps administrator prevent in-flight message processing from being interrupted by instance termination?

The correct solution is to enable instance scale-in protection when a message starts processing and disable it when processing finishes. This prevents the Auto Scaling group from terminating an instance while it's actively processing a long-running message, ensuring the task completes without interruption. Scale-in protection specifically prevents an instance from being terminated during a scale-in event. Changing the termination policy to OldestInstance or OldestLaunchConfiguration does not prevent the termination of an instance that is actively processing a message; it only changes which instance is terminated first. Suspending Launch and Terminate scaling processes for a specific instance is not a valid or practical approach for individual instances within an Auto Scaling group; these processes apply to the group as a whole.