Keeping the boxes honest: Portainer, Unraid and Glances from your pocket

The services you interact with tell you when they are unhappy. The machines underneath them mostly do not. A container that died at 3am, an array running in a degraded state, a host quietly pinned at full load, none of these announce themselves. You find out when something you do care about breaks downstream.

That is why a light monitoring habit pays off. Not a wall of graphs, just a few honest signals you can check in seconds.

Containers: Portainer

If you run things in Docker, the question is simply which containers are up and which are not. Portainer puts that in one view. A container that keeps restarting, or one that has been down since last night, is worth catching early rather than discovering when the thing it powers stops responding.

The array: Unraid

For an Unraid box, the signal that matters most is disk and array health. A drive throwing errors, or an array running degraded, is the kind of thing you want to know about now, while it is one problem, not later, when it has quietly become two. A glance at array status and disk health covers most of it.

The whole machine: Glances

Glances gives you the quick vitals of any host: CPU, memory, disk, network, load. It is read-only by nature, which is exactly right for a check-in. You are not trying to fix the box from a summary screen, you are trying to notice that it needs attention.

The habit, not the tooling

The specific tools matter less than the routine. Once every day or two, or whenever something feels off, run down the short list: are the containers up, is the array healthy, is anything pinned or out of room. It takes fifteen seconds, and it turns the big ugly surprises into small manageable ones. A homelab you glance at occasionally is far less stressful to own than one you only look at when it has already fallen over.