In 1.39 I was able to see Kubernetes pod level metrics (e.g. cpu, mem at the pod level). This seems to have gone away with 1.40 and 1.41.
I see a bunch of INFO level logs that mention all the pods but i don’t see anything in the dashboard. Is there a new configuration label that i need to set to true?
@data-smith : Thanks for reaching out to us.
I took a look at our code and also the dashboards and you should be able to see the pod level resource utilisation metrics (cpu, mem etc) under the Kubernetes State. Here is a snapshot from the latest agent nightly.
There is no additional configuration needed. @ilyam8 : Can you please confirm?
The image below shows what i’m talking about. You can click on a pod on the right and see cpu,mem and net info for it. It worked in v1.39. I’m not using Netdata Cloud. I’m running Netdata in a pod in Kubernetes.
We have released the V2 dashboards (as default) with the V1.41 and you can slice and dice on the k8s_state.pod_cpu_requests_used chart ( We have aggregated charts now and will not show up as individual charts (except on the old single node dashboard which should be eol soon).
Can you please confirm if you are using the new dashboard or the old one? At least these charts should still be visible as nothing has seemed to change from the collector.
In my Sections and Chart section I see Kubernetes Containers not Kubernetes State. And I don’t see any metrics for individual pods. I would give a screenshot but it’s on a private network and I don’t have permission to download things from it.
Menu items are the following:
…
Users
Anomaly Detection
Containers and VMS
Kubernetes Containers
Netdata Monitoring
@data-smith : Now I guess I know what’s the problem. Can you check if you have a node in your node list related to k8s-state? This is the agent that runs on the controller and provides metrics related to Kubernetes State.
Did you see any error during your deploy using helm?
Yea it’s running but i just realized it’s v0.2.7. I should upgrade. That should help.
Yes, that should do the trick. Let us know if the issue persists.
1 Like
Thanks for your help. I think my main issue is that you click the tabs like ‘instances’ to see all the metrics. Big UI change.
This UI change is intended for infrastructure monitoring where you want a view of your infrastructure (100s of 1000s of servers and applications) with a single pane of glass and where you cannot open each dashboard and monitor individually.
You can still use the old single node dashboard if you prefer it by selecting that option while opening your dashboard.
