LEGEND· 2017

Release Roulette · Medium · +50 XP · ~3 min

LEGEND · 2017 · AWS S3 us-east-1
Class: ops playbook misuse
Blast radius: half the internet

AWS S3 2017 — what should they have done?

Casualty stat: ~4 hours down · S3, EC2, console, half the web

Given what was visible at 09:37 PT, what was the right call?

Live signals

Playbook command intent

remove handful of capacity servers

What was actually entered

an argument that matched a much larger set

Validation on the tool

none — engineer responsibility

Time pressure

low — routine debugging

Engineer's recent context

deep in S3 internals for hours

Make the call

Playing anonymously. Sign in to save XP, streak, and badges.