This message was deleted.
# k3s
a
This message was deleted.
c
prometheus + grafana i kinda the state of the art
but it is not light on memory
q
I have a homelab of raspberry pis, so I've long looked for a "low memory" monitoring solution. As far as I've found out, such a thing doesn't exist. ๐Ÿ˜ž People making/designing monitoring solutions are usually the same people who monitor big professional clusters, where "low memory" is a distant memory of something they had to deal with 20 years ago ๐Ÿ˜› I've ended up using Grafana Cloud free tier (https://grafana.com/pricing/) to move a big part of the monitoring stack out of the cluster, that helped a lot. ๐Ÿ™‚
๐Ÿ‘ 1
b
Somewhat late to the party... we have Grafana Cloud in place on the RKE2 clusters, where Helm installs agents onto the worker nodes. Is it proper/safe/insane to want to try installing onto the local K3s cluster, which just has control plane nodes?