Developers run long CPU-heavy jobs on EC2 and frequently forget to terminate instances when idle. You must monitor CPU utilization and automatically terminate underutilized instances. Which solution meets this requirement?

The correct solution leverages EventBridge to react to Trusted Advisor findings. Trusted Advisor's "Low Utilization EC2 Instances" check specifically identifies instances that are underutilized. EventBridge can then be configured to listen for these specific Trusted Advisor events. When such an event is detected, EventBridge can directly invoke the EC2 TerminateInstances API as a target, automating the termination of idle instances. GuardDuty is a threat detection service, not designed for monitoring routine resource utilization, making it unsuitable. AWS Health Dashboard provides information about AWS service health and scheduled events, not individual EC2 instance utilization metrics. While Trusted Advisor can publish to SNS, directly using EventBridge to trigger the API call is a more streamlined and direct approach for this specific automation.