Capture generic process metrics (CPU usage, memory usage, i/o stats)

[grcevski]

This is similar to what happens today in Java/.NET runtime dashboards.

Stretch goal: Extend this to language specific runtime metrics, e.g. JVM/.NET GC, Go GC/Goroutines, NodeJS event queue length

Tasks

Preview Give feedback

1. Foundation to report application processes #892
   +0
2. Prometheus process metrics exporter #907
   +0
3. Report all the OTEL metrics in https://opentelemetry.io/docs/specs/semconv/system/process-metrics/
4. Report CPU metrics per user/system/wait state.
5. Decorate the process metrics with application metadata (service/ns name, k8s metadata, etc...)
6. cleanup/improve the code imported from NR