I’ve been setting up and testing prometheus and grafana for about a week now, since that seems to be the universally accepted solution for self-hosted monitoring. But I’m starting to question why it is so accepted. On top of prometheus not seeming useful on it’s own (needing grafana to visualize and alertmanager for alerts) it feels like with each thing i want to monitor I have to spin up another docker container to export/gather the data. There are other options like LibreNMS that seems to have all that built into one container. So what does this Prometheus/Grafana stack have that other monitoring services don’t? Is it really worth having to set up each of these specialized exporters and dashboards? Or am I mistaken that it’s the main solution everyone uses? Are you using something different for monitoring?
Grafana and Prometheus are great if you have numeric things you want to monitor. CPU usage, RAM, disks, throughput, etc. You can then do lots of things with these numbers, mainly compare them to your other systems or alert when they go out of bounds.
However, I very much prefer Zabbix for my home network monitoring as this is not so fixated on numbers but can easily work with e.g. error messages in logfiles and alert on those. Or I can regularly check a website for new firmware versions and alert once the latest version changes. There are also lots of ready-to-use templates available from their Community Hub.