A logistics company has been on AWS for several years and suspects its account contains resources that are provisioned but no longer doing useful work, such as unattached Elastic IP addresses, idle load balancers, low-utilisation EC2 instances and unassociated EBS volumes. Leadership wants a recurring, low-effort way to surface these specific kinds of waste so the team can decommission them, and they would prefer a managed feature over building bespoke utilisation queries. Which AWS capability is the BEST fit for routinely identifying this unused and idle infrastructure?
- AQuery the AWS Cost and Usage Report in Amazon Athena to list every resource whose hourly cost is greater than zero, treating the appearance of any charge as the signal that the resource is unused and should be removed.
- BEnable AWS Config with managed rules for each resource type and write custom remediation that infers idleness purely from whether a configuration item has changed during the most recent thirty-day evaluation window.
- CReview the AWS Trusted Advisor cost optimisation checks, which flag unassociated Elastic IPs, idle load balancers, underutilised EC2 instances and unattached EBS volumes for the team to act on each cycle. Correct
- DSet a billing alarm in Amazon CloudWatch on the EstimatedCharges metric and treat any month where total estimated charges rise above the previous month as confirmation that idle resources have accumulated in the account.
Why A is wrong: A nonzero charge proves a resource exists and costs money, not that it is idle, so this query would flag every active resource in the account and gives no signal about which ones are genuinely unused or underutilised.
Why B is wrong: AWS Config tracks configuration state and compliance, not utilisation, and a resource can be fully in use yet have an unchanged configuration, so config drift is a poor proxy for idleness and this approach is far more custom effort than required.
Why C is correct: Trusted Advisor's cost optimisation category includes managed checks for exactly these idle and underused resource types, so it routinely surfaces the specific waste leadership described as a managed feature, with no bespoke utilisation queries to build or maintain.
Why D is wrong: The EstimatedCharges billing alarm tracks total spend and cannot point to individual idle Elastic IPs, volumes or load balancers, so a rising bill might equally reflect legitimate growth and tells the team nothing about which resources to decommission.