2 comments

[ 3.4 ms ] story [ 13.8 ms ] thread
I use Netdata for a few homelab servers in lieu of learning and cultivating a complex grafana/prometheus setup (one docker container with a cut/paste token api vs a needy collection of interlinked pet containers with database config).

It’s great for things like “I just added a new docker container how does that affect cpu/memory over time” and some interesting (but inessential) automated alarms eg I get emails about Ethernet fifos along with advice on which Linux kernel options to fix it (which I ignore but it’s fascinating).

Doubt I’d ever need more than the 5 servers the free plan offers but even if I did I would not pay the $90 per year home lab license fee.

One thing stopping me deploying netdata on more servers is that it is very resource intensive in default config. I tried it on a low power celeron server and my ancient Synology and in both cases the agent was consuming 10-20% CPU before I nuked it. I think you can lower the polling but I didn’t spend much time investigating.

For the technical issues you mentioned, please consult our guide here: https://learn.netdata.cloud/docs/netdata-agent/sizing-netdat...

As for our subscription plans, if you're managing fewer than 6 nodes, you can always be on our Community plan. Should you have more nodes and prefer not to choose one of our Netdata Cloud plans, you can utilize our agent dashboard.