Recreate BigBang

Take everything down, deploy the new version everywhere, bring it back up — simplest possible strategy, but means full downtime and full blast radius if it fails.

Why we need this / what value this brings

Simplicity has a real cost here: it maximizes blast radius, since there's no staged subset to catch a problem before it hits everything.

When to use this

Low-stakes changes, or environments where partial-version-mismatch during a staged rollout is worse than brief full downtime.

How to use or implement this

If used, at minimum make sure verification runs immediately post-deploy with an automatic halt/alert on failure, since there's no smaller blast radius to fall back on.

Research questions

• Was the original golden-image push effectively a big-bang deploy across the fleet, or was it already batched/staged?

---

Empty folder — drop notes, links, and findings here as you research.