A popular web application runs on multiple Linux EC2 instances in an Auto Scaling group inside a private subnet, with an Application Load Balancer targeting those instances. Systems Manager Session Manager and SSM Agent are configured on all instances. After a new release, some instances are marked unhealthy and are terminated, reducing capacity. CloudWatch logs are inconclusive. How should you obtain access to an instance for troubleshooting?

Suspending the Auto Scaling group's Terminate process prevents instances marked unhealthy from being terminated, allowing you to connect via Session Manager for troubleshooting. This ensures the problematic instance remains available for investigation. Suspending the HealthCheck process would prevent the Auto Scaling group from identifying unhealthy instances, which is not the goal. Enabling EC2 instance termination protection is a good general practice but doesn't prevent the Auto Scaling group from attempting to terminate an instance based on its policies; it just prevents manual termination. Setting the termination policy to OldestInstance changes which instance is terminated, but doesn't stop the termination of unhealthy instances.