Kill the staging-server tax

Most teams pay for production, then pay again for staging that looks roughly like production, then pay a third time for QA that's drifting from both. Three EC2 instances, three RDS bills, three sets of monitoring agents. The staging tax is real money for an environment used twelve hours a week.

On Hoody, staging is a snapshot of the production container. It sits on the same bare metal as production, restored from a snapshot whenever someone needs it, frozen back to disk when they're done. No second instance, no drift, no duplicate database to keep in sync.

QA gets its own snapshot too, branched off staging at known-good. Three environments, one machine, one bill. Drift is impossible because they all started identical.

Staging used to be a duplicate of production. Now it's a snapshot of it.