![[?]](https://social.dk-libre.fr/social/oldsysops/s/7e93fa6bf16b47da4f57ddf9072aff8c.png)
...
![[?]](https://files.mastodon.social/accounts/avatars/000/035/310/original/43aec69802c8a4c9.jpg)
@oldsysops @Anarcat @noodles I will cough awkwardly in the corner. I built our Prometheus at work (and run it on my own machines), but I feel it's definitely not easy to set up if you want alerts or certain sorts of checks (like 'is this port responding' or 'can we resolve this DNS name' or 'are we seeing an all-is-okay response from a health check'; for Prometheus people, Blackbox indirect scrapes and modules and etc etc etc).
It may still be the best option but be ready for a learning curve.
...
![[?]](https://kolektiva.social/system/accounts/avatars/000/004/395/original/123051a70a5f6fd8.jpg)
@cks @noodles @oldsysops if you want to do EVERYTHING, no, it's not trivial, but then you're asking to solve a non-trivial problem... that said, the blackbox exporter pattern is particularly annoyingly hard
@cks @noodles @oldsysops to put this another way, if you start from scratch again now, would you do anything different? i wouldn't.
...
@Anarcat @noodles @oldsysops My view is that for work-scale problems, anything as good as Prometheus is going to be about as complicated because it's a hard problem. Plus you get a lot of value out of Grafana being able to talk to Prometheus (or any other equally complex metrics source).
But for simple one-machine use it feels like overkill. I hope there's something simpler and more basic out there.
(I have it on my own machines because I can copy the work config and use them for testing.)
![[?]](https://social.earth.li/media/42c31acd40ce187d188961f93891e5a88990a315166525d226de40ce5650ed8f.png)
...
@noodles @oldsysops @cks this sounds just like photography, where "the best camera is the one you have with you when you need it"