I'm slowly working on upgrading my monitoring stack. My criteria are that the software must have a small enough codebase that I can eventually understand it, a low memory and CPU footprint, and be convenient to use. Being written in Go, C, or shell scripts is a plus since these are the languages I primarily use, but it's not strictly necessary.
During this journey, I ended up writing a tool to monitor the availability of websites. It’s just a small part of the stack, but an essential one. I know there are millions of similar tools out there, but I wanted to create one that fits my own preferences. The most important aspects for me were ease of configuration, keeping the criteria I mentioned earlier, and doing exactly what it’s supposed to do, monitor site availability, and nothing else.
It exposes metrics via files, with the idea that another daemon can pick them up and, for example, write them to a database or display them in a dashboard with graphs. I really like the idea of exposing internal data of daemons through files, similar to how Linux uses the /proc directory, so that’s my usual approach for long-running processes.
I'm adding features as they become necessary. I have a few ideas for the future, but I’ll only implement them when they truly become useful to me.
If you’re interested, you can check it out here:
[https://github.com/thekhanj/avail
](https://github.com/thekhanj/avail)
Comments (0)