Improve documentation for controller manager metrics

[constanca-m]

Problem

When users try to monitor their Kubernetes application, they don't have a way to understand our predefined dashboards without having solid knowledge about Kubernetes.

Right now, the documentation for that can be found in this page. However, this is not enough to understand all the metrics and the meaning of each one in here doesn't it make it clearer.

Solution

Create a new documentation that highlights the metrics available on the controller manager dataset and which problems we can detect with those. With this new document we can expect to have the answer for these type of questions:

• What can we conclude from the unfinished work of some workqueue? What does that mean?
• Why is it important to track the CPU time spent?
• What is the take away on resident memory variations?

This new page would use visualizations to make the answer for these type of questions more clear. This page should not use the predefined dashboard, as a change on it would require an update to this page.

[gizas]

Constanca do you think this issue and this elastic/integrations#4886 can be combined into one?

[constanca-m]

I think so @gizas , but since we haven't decided on a common structure, I think it is better to start with just one thing and see if it is worth it.

[rameshelastic]

@constanca-m , is this issue being worked on?

[constanca-m]

It is not @rameshelastic. I am only working on TSDB now