@jonty and this particular instance is not actually even a big deal it just didn't manage to do a periodic scrubbing time because it was too busy shuffling data amongst itself, it's just that ceph doesn't really have "really big problem" and a " minor problem " difference...

Then again, the last piece of software that I used that had that (green, orange, red) was elastic search and honestly speaking that traffic lights system status is burned right into my head as something they can go straight to hell

I guess error reporting is just hard, every single data center hall I have walked through had every single router in at least some state of major/minor alarm state. My favorite is the equipment has a dedicated " catastrophic " alarm state.

@dwm @jonty Unfortunately I've discovered many ceph warn state's that are actually "I am not accepting writes anymore lol", something that is generally fatal to quite a lot of workloads

@hugh @bix @jonty as a person in a rank near me who currently has a broken power supply (or at least for the last 3 years they have) and I've been extremely tempted to order the spare part and just install it for them because I'm so annoyed of the extremely loud beeping