| 24 Feb 2025 |
 K900 | * The Prometheus metric for "entity available" returns 0 if the entity is in state unknown | 17:04:02 |
| K900 | Except like | 17:04:11 |
| K900 | Most buttons default to state unknown | 17:04:15 |
| K900 | And a bunch of integrations report their state as unknown | 17:04:47 |
| K900 | And there's literally no way on the Prometheus side to tell if it's unknown or unavailable | 17:04:57 |
| K900 | And here I was thinking that I should just set up an alert for when a device becomes unavailable HOW HARD COULD THAT BE | 17:05:12 |
| K900 | TURNS OUT VERY FUCKING HARD AND IT'S NOT EVEN MY FAULT | 17:05:23 |
| K900 | @hexa what are my odds of getting this fixed | 17:06:47 |
| K900 | If I make a nicely written bug report about it | 17:06:53 |
 @hexa:lossy.network | fairly good if you file the issue first | 17:11:20 |
| @hexa:lossy.network | and submit a fix to close that issue | 17:11:27 |
| K900 | Problem is, I can' | 17:11:38 |
| @hexa:lossy.network | don't mention using nixos 😛 | 17:11:40 |
| K900 | * Problem is, I can't really think of an immediately obvious fix | 17:11:45 |
| K900 | But I will file an issue | 17:11:53 |
| @hexa:lossy.network | then mention the prometheus component maintainer | 17:12:12 |
