Mhm, anyone knows a minimalistic monitoring solution that doesn't require me to keep a zoo of daemons and services around?

(And that is agent-based so I don't need to expose an API on each of my servers for no particular reason 👀)

@sheogorath #netdata?

I don't know why you have a problem with pull-based monitoring though.

I love #netdata. the only put-off I have with it is that it is agent-based. I guess it works for you.

@sheogorath Glances should be an option, but if you are not allergic to #webuifromhell, I'll advice xymon

Which is very simple, quick to use, flexible (a custom check is just some lines of bash, python, or every preffered language) and despite its first look, very clear to use with the red/yellow/green color ^^

@sheogorath If it's just about collecting (not notification) then it sounds like writing to RRDTool might be an option.

@menn0 i can also recommend Telegraf. It has Plugins for pretty much anything, or you can just read json from somewhere or you can execute commands and read its output... And so on.

@blacklight447 no, I have status cake for that. Looking more towards system metrics and more centralised logging.

But I think I found a sufficient setup in a TIG stack for now.

@sheogorath I'm curious to know why you decided with Grafana against Kapacitor. Why not the integrated #TICKstack from #InfluxData? I'm sure that would provide better integration together.

@shine one major concern of mine here is to keep the toolchain small. There is no need for me to have everything isolated such as kapacitor and chronograph. I don't want to maintain this zoo.

All the setup needs to monitor is less than 5 machines and it's not expected to grow anywhere. So scaling is not a concern. And this means the microservice architecture just costs me time in configuration, debugging and upgrading.

Sign in to participate in the conversation
Sheogorath's Microblog

This is my personal microblog. It's filled with my fun, joy and silliness.