Checkpoint: Release Engineer
Prove the production instincts: environment discipline, release strategy selection, incident triage, and honest health checks.
This checkpoint grades release engineering judgment in runnable form — the decisions that separate "it deployed" from "it deployed safely and we can prove it." ...
Shipping as a discipline
Deployment is where your abstractions pay rent: env vars earn their indirection, health checks earn their uptime, structured logs earn their 11pm usefulness. The checkpoint asks you to decide, not recall — because on call, nobody asks you what a canary is; they ask what you are going to do in the next ten minutes.
The deploy is not the finish line
Real release engineering is what happens around the deploy: config that fails fast, migrations that are reversible, health checks that gate traffic, logs you can search at 2am, and a rollback you can run before the incident channel finishes typing. The checkpoint's scenarios hand you exactly those moments — your answers are the muscle memory the next on-call shift will borrow.