A company migrated to AWS and has AWS Business Support. It wants to monitor cost-effectiveness of Amazon EC2 instances across accounts. Instances are tagged by department, business unit, and environment. Development instances have high cost but low utilization. The company needs to detect and stop underutilized development EC2 instances. An instance is underutilized if it had average daily CPU utilization of 10% or less and network I/O of 5 MB or less for at least 4 of the past 14 days. Which solution achieves this with the LEAST operational overhead?

The correct answer leverages AWS Trusted Advisor, which automatically identifies underutilized EC2 instances based on predefined criteria, including CPU utilization and network I/O, aligning with the problem's requirements. EventBridge can directly detect these Trusted Advisor findings, triggering a Lambda function to apply the specified tag filters and stop the instances. This approach has the least operational overhead because Trusted Advisor handles the complex utilization analysis and reporting. Incorrect options: Creating CloudWatch dashboards requires manual configuration of metrics and alarms for each instance or group, which is more operationally intensive than using Trusted Advisor's built-in analysis. AWS Systems Manager does not inherently track EC2 utilization in the way described; CloudWatch is the primary service for metrics. This option introduces unnecessary complexity. Building a custom Lambda function to retrieve and store metrics in DynamoDB and then using QuickSight for analysis is a highly custom and operationally intensive solution, requiring significant development and maintenance.