Notes from a staging cleanup

Every few months we set aside a day to go through client infrastructure and ask a boring question: does anything here exist that nobody can explain?

Staging environments are the usual answer. They get spun up for a specific migration or a specific demo, they do their job, and then everyone moves on to the next thing without circling back to turn them off. Six months later there's a box somewhere, still reachable, still running whatever configuration it had on the day it was last useful, with credentials nobody's rotated because nobody remembers it's there.

None of this is exotic. It's the same handful of causes every time: a deploy that was meant to be temporary, a DNS entry that outlived the thing it pointed to, a config file that got copied along for convenience and never cleaned up. The fix isn't clever either — it's just discipline. When an environment's job is done, it comes down. Not archived, not "paused," down.

We'd rather write this kind of thing into a client's process up front than find it during an audit later.