1
Choose a realistic failure
A slow dependency or a lost zone beats a made-up scenario.
⚠ Guide
Incident response drills your on-call will thank you for
An incident response drill is a rehearsal where you break something real, page the on-call team and run the full response, roles, channel, status page, while the stakes are zero. Unlike a tabletop exercise, it tests whether alerts fire and runbooks work, not just whether people can talk through them.
Run a drill in five steps
2
Decide who knows
Surprise drills train detection. Announced drills train the fix.
3
Inject and page
Let alerts fire through the normal path. Do not shortcut paging.
4
Run it like it's real
Incident channel, roles, status page draft, customer comms.
What a good drill reveals
Alerts that go to someone who left. Runbooks with missing steps. Nobody sure who owns customer comms. These are cheap to fix after a drill and expensive to discover during an outage.
Incident drill FAQ
+ How often should we run incident response drills?
Monthly for on-call rotations, so every engineer practises at least once a quarter.
+ Should drills be a surprise?
Mix both. Start announced so people learn the process, then add surprise drills to test detection.
Safe by design
Every fault Oh My GDaaS runs is aimed at infrastructure you have proved you own, through a DNS record, a scoped cloud role or a deployed agent. Every run has a blast-radius limit, a hard time limit and an abort button. We are not a booter, stresser or DDoS service and never will be.