[justusthane]

Apologies for the hijack, but I’ve been looking for a particular kind of monitoring tool lately that I’m not sure exists.

I would like something that allows me to write my own arbitrary monitoring scripts in whatever language I want, and the tool would take care of everything else: scheduling and running the scripts, parsing the output, alerting, authentication, presenting the info on a pretty dashboard with graphs, etc.

I think Monit can do this to some extent, but I haven’t explored it yet — it looks like the dashboard and info presented is a lot simpler than what I’m looking for.

Is there some reason this isn’t a useful concept? For context, I’m looking at this from a homelab/selfhosting/hacking perspective.

[moehm]

Good ol' Nagios does this, except maybe for pretty graphing. Maybe look into Icinga2 (fork of Nagios) for that.

The modern Version would be something like Prometheus with Grafana.

Zabbix might be too much for a homelab setup.

[mbirth]

Actually, Zabbix runs pretty well in a homelab setup. Doesn’t eat much resources with its 3 containers here.

[supriyo-biswas]

I’m not sure how useful this is from a homelab setting as the services are private, but what I’ve typically done to address this need is to write AWS Lambda functions, make them accessible over HTTP, and have the uptime monitor (I personally use uptime kuma) monitor that HTTP endpoint. I can then return a 4xx or 5xx response from the function when a certain condition isn’t as expected.