`kubernetes.volume.fs.used.pct` has incorrect type in index template

Elastic Stack Beats
,
1

It seems that field kubernetes.volume.fs.used.pct has an incorrect type in the index template metrics-kubernetes.volume (automatically generated by metricbeat or fleet).

The type specified by the template's mapping is long, while it is explicitly specified in the Metricbeat exported field specification that the field should be scaled_float.

The mapping specified in the metrics-kubernetes.volume index template:
image

The Metricbeat specification about this exported field:
image

It doesn't make sense to use a Long as the values sent by Metricbeats are 0-1 range percentage, which leads Kibana to round all percentages to 0%.

image
image204×526 3.62 KB

I'm using managed Elastic Fleet 7.15 and I retrieve Kubernetes metrics using the Kubernetes integration (1.1.0) from Elastic Fleet.

Did I dramatically misunderstand the situation or is there a bug here? If this is a bug, do you see any solution for me to build charts based on volume percentage alternatively?

Kind regards

2

Hi lepouletsuisse! Thanks for pointing this out.
This is indeed a bug and will be fixed in Kubernetes Integration 1.5.1 .
kubernetes.volume.fs.used.pct is the result of kubernetes.volume.fs.used.bytes divided by kubernetes.volume.fs.capacity.bytes.

So meanwhile you can create a custom chart where you can represent the percentage with a Formula like

median(kubernetes.volume.fs.used.bytes)/median(kubernetes.volume.fs.capacity.bytes)
3

Thank you very much!
I'll try your solution to do my chart.

4

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.