I need metrics and graphs and maybe logs for some servers I run on a personal project.
At my day job this would mean Grafana or ELK.
I want something significantly less heavy than that. I want to run logging and metrics visualization on the same server I run production workloads, and I don't want the metrics to drag down those production workloads.
What's a good option here? Or am I stuck with the Enterprise options?
@ajroach42 I used to run (or still do on my homeserver, but didn't even check if it still works and am going to deprecate it) collectd with rrdscout (not maintained anymore) as frontend.
My more contemporary monitoring deployment is prometheus + grafana.
And I have the impression that prometheus itself isn't just much more powerful with promql, but also okay resource-wise. I'd still like to have something lighter than grafana as frontend, but haven't found anything better either. :/
@phryk Yeah. I figured prometheus and grafana is where I would land.
@ajroach42 @phryk I use InfluxDB and Grafana for my homelab, on the same box that runs everything else. It doesn't interfere with anything. Box is a ThinkCentre M910q Tiny (7th gen core) I scored for $100 off local Craigslist.
I don't recommend InfluxDB, I'd use Prometheus these days too. Grafana is good.
@ajroach42 That said, if you don't need something as versatile as prometheus, collectd might be worth a look.
Outputs RRD, so there's a ton of possible frontends, but I've never quite found "the one" for me.
Added dark look SVG output to RRDScout, but that was not exactly resource friendly…^^
Alternatively, munin – has back- and frontend integrated.
But In the end, prometheus+grafana is just in another league – the first monitoring setup that can actually scratch my various itches.
𝚜𝚎𝚕𝚎𝚊 
@ajroach42 https://oss.oetiker.ch/rrdtool/ might be of interest. For an example see: https://aplawrence.com/Girish/gv-rrdtool.html
@ajroach42 Prometheus has typically been pretty lightweight in my experience, and Grafana pointed at it has very little server-side resource requirements.
@aschmitz That's good to know.
@ajroach42 At home I'm using Zabbix.
I'm using netdata for now. It's working.
This server is Chugging.
Poor thing.
I guess I shouldn't be surprised. I asked it to do a lot.
Time to migrate some workloads to another box.