Retrieval of memory and CPU usage metrics in hardened Kubernetes cluster

[Vincehood]

Hello,

I am after a solution allowing to monitor the CPU and memory usage of containers deployed within my application namespace on top of a hardened Kubernetes cluster.

The solution shall not:

- leverage on any cluster role (scraping of kubelet metrics is therefore excluded)

- assume that a Prometheus federation endpoint is exposed by the hardened Kubernetes cluster

Is there any solution available (or in the works) to the needs above?

Thanks