Prometheus experience

larsbnd

Hi folks, does anyone have Prometheus along with Enhance running? Hence Prometheus is the go to multiserver environment and especially as a container monitoring tool I wondered if someone uses it and if anything out of the usual was a difficulty or so.

Would love to hear some feedback.

MediaServe

I gave NetData a try and was quite disappointed. The load generated by the agent was too much, and the ability to delete a node was extremely difficult.

Enhance doesn't "dockerize" websites (or other roles) either. Just the primary platform containers. I'll definitely be interested in Prometheus but don't really want to spend time without some feedback, so I'm registering my interest too!

Currently I use an internal server monitoring system that works well for cPanel but not as much with Enhance other then alerting for basic outages. I've been using https://github.com/jesseduffield/lazydocker for docker stats but the difficulty is monitoring non-docker containers.

Adam

MediaServe Enhance doesn't "dockerize" websites (or other roles) either.

Just to clarify, the websites are containerised in the same way as Docker, it's just done within Enhance code rather than Docker for better performance. It's unlikely that any 3rd party tools will be able to monitor these however we are adding CPU, memory and IO stats and graphs in an upcoming update.

xyzulu

MediaServe The load generated by the agent was too much

I estimate it adds about 5% to loads generally.

I'm interested to hear what you are doing. For me, it's not "live" issues that I have an issue with.. I remotely monitor and have no trouble finding an issue while it is happening. The thing that netdata allows me to do it look back in time and see how things looked then.

Lazydocker? Are there any kinds of history stats using that? Even past 24hrs is enough for me.

MediaServe

Adam Right, I could have been more clear that Enhance still creates their own containers for websites and other roles.

It's unlikely that any 3rd party tools will be able to monitor these however we are adding CPU, memory and IO stats and graphs in an upcoming update.

This will be greatly appreciated! I hope these can be provided at the website, server and estate level sorted from largest resource consumers to least.

8Dweb

MediaServe This will be greatly appreciated! I hope these can be provided at the website, server and estate level sorted from largest resource consumers to least.

Yes, please.

larsbnd

For Prometheus and Grafana it's just convenient to scrape different endpoints and have them sorted in one place and get automated alarms about any sort of thing. I am personally utilizing crowdsec and its handy to have endpoints for Prometheus right out of the box: https://docs.crowdsec.net/docs/next/observability/prometheus/