A city FinOps team reviews Cost and Usage Reports for sandbox OUs and finds Classic Load Balancers with zero traffic for months plus many unattached EBS volumes. Leadership wants a durable process to surface these waste patterns before renewing the next budget cycle. Which approach best identifies the unused resources from usage evidence?

For a city FinOps team cleaning sandbox organizational units before the next budget cycle, unused-resource discovery must come from usage evidence in Cost and Usage Reports or Cost Explorer, not from folklore about which Classic Load Balancers look old. Query load-balancer and EBS line items for near-zero traffic and unattached volumes, then target idle Classic Load Balancers and orphaned disks in those sandbox accounts for governed cleanup. That approach turns the municipal meter tape into a durable waste-detection process leadership can repeat each appropriation cycle. Deleting every Classic Load Balancer in production because the product is legacy fails on concept: legacy status is not proof of zero traffic, and production citizen paths may still depend on those balancers until a measured migration completes. Relying only on monthly invoice PDF totals fails because PDFs lack service-level and resource-level dimensions needed to distinguish idle balancers and unattached volumes from healthy spend. Moving all sandbox accounts behind a single shared NAT Gateway fails because consolidating NAT does not remove idle Classic Load Balancer hours or orphaned EBS charges from CUR; it merely rearranges networking line items. Exam caveat: SAP-C02 often tempts candidates to equate 'legacy' with 'delete now' or to treat invoice totals as FinOps granularity—prefer usage-backed identification before cleanup. Operational check: export CUR or Cost Explorer filters for CLB and EBS with near-zero metrics in sandbox OUs, ticket owners for confirmation, delete or stop under change control, and re-measure the same filters next month so the budget office sees waste trending down before renewal.