In my day job I kept noticing CloudWatch alarms that were technically there, but in practice didn’t help much anymore - no actions, disabled actions, or alarms sitting in ALARM / INSUFFICIENT_DATA for weeks that everyone just got used to.

Once you have a lot of services and a lot of people touching things over time, it gets surprisingly hard to tell which alarms still make sense.

I ended up hacking together a small, read-only CLI that just lists alarms across regions and highlights those kinds of cases.

This isn’t meant to fix anything automatically — mostly a way to see what you actually have. Curious if this matches what others see in real setups, or if you approach this problem differently.

Repo: https://github.com/wrybakiewicz/cw-alarm-audit