Kubernetes monitors should use the relabel component internally instead of of their own implementation

[mattdurham]

Request

The kuberenetes monitors have relabel rules but do not use the caching in the relabel component which can create issues with CPU.

Use case

Reduce cpu

[csh0101]

what's this problem means?

[mattdurham]

With an extreme number of labels the CPU will skyrocket instead of using the relabel component internally that has a cache to handle this situation. In reviewing some issues we found that there were high CPU spikes whenever service monitor went to relabel that would have been addressed with using a cache.