our staging environment has drifted so far from prod that I've basically stopped trusting it. it happened gradually -- a config change here, a manual fix there, someone needed something 'just for testing' and now the state is a mess.
I've seen people suggest infra-as-code, automated deploys to staging on every merge, scheduled resets, etc. but in practice I'm curious what actually works for teams under like 10 engineers where nobody owns DevOps full time.
what's the lightest-weight solution that actually keeps staging usable?
[link] [留言]