A healthcare SaaS must prove its ransomware runbook for EC2 and RDS actually restores service within four hours, but leadership forbids disrupting production. The plan requires isolated testing of containment, backup restore, and failover steps with measured results. Which procedure validates the effectiveness of this incident response plan?

Game-day simulations with Fault Injection Service in an isolated environment validate the full response sequence: teams can rehearse containment, restore from clean backups, and fail over while Resilience Hub measures recovery against the four-hour RTO. Fault Injection Service cannot encrypt data itself, so the ransomware effect must be staged; isolation lets you test the runbook without touching production. AWS Backup automated restore tests in an isolated VPC demonstrate that EC2 AMIs and RDS snapshots are restorable and provide timing evidence, but they validate backup recoverability only, not network containment or failover orchestration. Fault Injection Service experiments in an isolated staging account can rehearse EC2 containment and RDS failover, building team readiness, but without Resilience Hub measurement they do not provide the required measured proof of the time-bound target. Isolated drills using Security Groups for containment and RDS Multi-AZ failover with Resilience Hub tracking provide failover metrics, but assume a healthy database and omit restoring clean backups—the decisive step when ransomware has encrypted data. As an exam caveat, a successful restore, drill, or detection activity alone is not full response validation unless all required steps and time-bound metrics are covered together. As a practical check, run the game-day in a dedicated test account with VPC isolation, pre-provisioned clean backups, and Resilience Hub configured for the recovery application before declaring the runbook proven.