Capture metrics per node using Prometheus

This issue was open based on #32266 (comment 236652404) in order to accommodate the "Metric visualization" #32266 (comment 235836441)

We need to automatically capture per pod metrics using Prometheus

Following metrics should be (we should add more if needed)

1. CPU
2. Memory usage
3. Disk i/o
4. network In
5. network Out

We should capture also advance metrics such as:

1. Desired pods
2. Available pods
3. Unavailable pods

If it's possible also to capture containers information I can specify what is needed, if not I can open a new issue

Designs