Fixed metrics path and deprecated labels for the cadvisor job

[ppatierno]

Fixed dashboards using the right label

Signed-off-by: Paolo Patierno ppatierno@live.com

Type of change

Select the type of your PR

• Bugfix

Description

This PR fixes #3308.
Memory and CPU usage metrics are not working anymore on Kafka and ZooKeeper dashboards due to change on the path where cadvisor exposes the metrics and some deprecated and dropped labels on the related metrics.
Following the current not working metrics for Kafka and ZooKeeper dashboards.

and here the working ones using this PR.

Checklist

Please go through this checklist and make sure all applicable tasks have been done

• Try your changes from Pod inside your Kubernetes and OpenShift cluster, not just locally
• Reference relevant issue(s) and close them after merging

[tombentley]

The referenced Kubernetes PR says

Duplicate labels were included in the 1.14 and 1.15 releases

Merging this PR would break support for Kubernetes 1.11–1.13. I know we're on the verge of dropping support for those versions ourselves, but we've not yet done so.

[ppatierno]

We have always said that they are examples right? We don't provide support for them so we are not breaking compatibility with something we don't actually support. What are you proposing? Having the metrics working only on older Kubernetes version and having them broken on new ones?

[tombentley]

I'd forgotten that we've always said these were examples, rather than anything supported. However, it's probably worth mentioning in the changelog that we're dropping support for metrics in those older versions, and linking to this PR, so that it's more clear that we've done this on purpose, and users on older versions at least know how to fix things.