Orphaned Staging RDS Instances Are Quietly Draining AWS Budgets

AWS cloud costs in engineering teams are often inflated by forgotten staging resources, particularly RDS database instances that remain active long after the feature branches they supported have been merged and shipped to production. Unlike EC2 instances, which have clear individual owners, staging databases are typically assigned to a team, meaning no single person feels responsible for decommissioning them. A typical pattern sees one engineer provision a database for a sprint, another later run tests against it, and the instance then linger for months with no one willing to delete something a colleague might still need. The problem is compounded by the fact that ECS services and Auto Scaling Groups tied to the same workflow also keep running, since no default AWS lifecycle policy stops them. The root cause is less a tooling gap and more a visibility and incentives failure, where cleanup costs are diffuse, ownership is unclear, and ignoring idle resources requires less effort than removing them.