What DevOps Engineers Should Know Before Running an AWS DR Drill
Lessons from leading a full AWS disaster recovery drill from Singapore to Bangkok, and why a test run first matters.
This is a short summary. Read the full article on LinkedIn →
I led a full AWS disaster recovery exercise, failing over from Singapore to Bangkok with an 18-hour RPO and a 4-hour RTO as targets. The single most valuable step was a complete test run two days before the real drill: it surfaced more than 20 issues that documentation alone would never have revealed.
Key takeaways
- Your DR region isn’t a copy of your primary: check instance availability, quotas, service-linked roles and SDK support before you fail over.
- Run a full practice drill first to catch infrastructure, configuration and operational gaps.
- Size databases for recovery speed, not just normal traffic; parallel restore tools cut restore times significantly.
- Isolate the recovered environment from the source by checking every shared dependency, such as queues and webhooks.