Disaster recovery (DR) testing is the practice of deliberately running your recovery plan, in whole or in part, to prove that systems and data can be restored within the targets you have set. A plan that has never been run is an assumption. AWS's Well-Architected guidance puts it plainly: testing is the only way to ensure a workload will operate as designed, and the tests should be exercised regularly.

Common types of test

TestWhat it provesDisruption
Walkthrough or tabletopPeople know their roles and the runbook makes senseNone
Restore testBackups can actually be restored, and how long that takesLow
Recovery drillA workload can be launched in the recovery location and usedLow if isolated from production
Failover testProduction can really fail over and, afterwards, fail backPlanned and controlled
Game day or chaos experimentThe system and team cope with injected faultsControlled, with stop conditions

What AWS offers

The Well-Architected Reliability Pillar also lists two practices by name: REL12-BP04, testing resiliency using chaos engineering, and REL12-BP05, conducting game days regularly.

The UK regulatory angle

For firms in scope of the FCA's operational resilience rules, testing is how you show you can stay within impact tolerances. The FCA says firms had until 31 March 2025 to be able to operate important business services within their impact tolerances, having carried out mapping and testing, and to be on track to remediate vulnerabilities that testing exposed. See FCA operational resilience.

Getting started

Test against the Recovery Time Objective you have set, time it, write down what broke, and fix it before the next run. Our guide to testing an AWS DR plan walks through a low-effort first drill, and our AWS disaster recovery service runs and evidences the tests for you.