An application runs on EC2 instances behind an Application Load Balancer and is scaled with an Auto Scaling group. Occasionally the application becomes slow; CloudWatch shows some instances have very high CPU. The administrator must build a CloudWatch dashboard that automatically shows CPU metrics for every EC2 instance, including instances launched later by Auto Scaling. What is the most operationally efficient way to do this?

The most operationally efficient way is to use CloudWatch Metrics Explorer to filter by the aws:autoscaling:groupName tag. This tag is automatically applied to instances launched by an Auto Scaling group. By filtering on this tag, the dashboard will dynamically include CPU utilization metrics for all current and future instances within that Auto Scaling group without manual intervention or custom scripting. Option 1 is overly complex and introduces unnecessary operational overhead with Lambda functions and subscription management. Option 2 requires custom development and maintenance of agents/scripts on each instance, which is less efficient than leveraging built-in CloudWatch features. Option 4 is incorrect because filtering by instance state (e.g., "running") would include all running instances, not specifically those from the Auto Scaling group, and wouldn't automatically include new instances in the same targeted way as filtering by the Auto Scaling group tag.