A legacy application on EC2 instances fails when total CPU utilization exceeds 80%. As a temporary mitigation while the app is rewritten, a SysOps administrator must restart instances when CPU goes above 80%. Which approach accomplishes this with the LEAST operational overhead?

The most straightforward and least operationally intensive solution is to use a CloudWatch alarm with a direct EC2 reboot action. CloudWatch can monitor CPU utilization metrics natively. When the threshold is breached, the alarm can be configured to directly initiate an EC2 instance reboot, requiring no additional custom code or complex integrations. Deploying a cron job with a custom script introduces significant operational overhead for maintenance, monitoring, and error handling. Configuring an EventBridge rule to invoke a Lambda function adds complexity with Lambda development, testing, and management. While using a CloudWatch alarm to trigger an AWS Systems Manager Automation runbook is a valid approach for more complex remediation, for a simple reboot, the direct CloudWatch alarm action is more efficient and has less overhead.