How to create a visualization for node cpu usage of kubernetes

[Kaushik_R]

Hi All,

I'm trying to create a visualization to capture the node cpu usage of a kubernetes node.

i'm trying to use kubernetes.node.cpu.usage.nanocores , kubernetes.node.cpu.usage.nanocores but none is working the outcome is 0

The aggregation that I tried using is average ,max and even unique count .

We are not using metricbeat module here but filebeat is configured.

Can anyone please help me how to achieve this ?

[devon.thomson]

I'm not sure, but it seems to me like there isn't any relevant data in your current view. The first step I would take is to set the time range to a range where you're sure there will be data.

If nothing shows up, I would look next in Discover to see if you can find any records at all, and if so, check if the field kubernetes.node.cpu.usage.nanocores exists in the records that are there. You can add a filter for that by selecting the field, and the operator exists.

If the data isn't there at all, it must be an ingestion problem, and I will forward you to the appropriate team!

[Kaushik_R]

Hi Devon,

Thanks for the reply .

I do see the data show up but as soon as I add another metric related to nodeGroup or node.label in order to group based on the nodes the data vanishes.

[devon.thomson]

You should check again in Discover and make sure there are records where nodeGroup or node.label exists, as well as kubernetes.node.cpu.usage.nanocores existing.

Are you using Visualize or Lens to create your visualization?

[ghudgins (Graham Hudgins)]

@Kaushik_R

We are not using metricbeat module here but filebeat is configured.

I don't think that field is included in filebeat so if you aren't feeding metric data that could be it.

...but this field is in metricbeat. I would double check your data in Discover like @devon.thomson was mentioning using an "exists" filter for the nanocores field and expand a single document from that dataset to get a sense for what's there. I don't see this nodeGroup or node.label as part of the event.dataset for "kubernetes.node" data on my example environment...here's what I see:

...but i'm able to build a visualization with the fields that are there:

[Kaushik_R]

Hi @devon.thomson and @ghudgins ,

Thanks for your support.

You both were right the metric is part of the metricbeat module my bad and in my case i was not able to see the data show up in the discover section.

Index : mon-metrics-k8s-*
Metric : kubernetes.node.cpu.usage.nanocores
Another Metric : kubernetes.node.labels.occloud_oracle_com/nodeGroup ( using this to group the value from spilt series option )

So this is the reason it's not coming up in visualization as well.

will you guys let me know what is the reason for data not coming up and Do i need to use some other index ?