Update config pattern of Stream pods to expect METRICS_PORT env var.
This value is then used to create the Prometheus metrics server. Metrics
server uses the global registry.

Update Operator to pass along necessary config for metrics ports &
update Stream StatefulSet pod spec to expose metrics port on the
container.

Add metrics instrumentation for:
- K8s resource watchers and Stream Subscribers.
- Stream current offset (counter).
- Stream subscriber count (gauge).
- Stream subscriber group last offset processed (counter).
- Per Pipeline last offset processed (counter).
- Per Pipeline active instances (gauge).
- Per Pipeline number of stage subscribers (gauge).

Added Prometheus metrics covering:
- Number of leadership changes.
- Leadership state (1.0 == leader, anything else is follower).
- Error counter for K8s resource watchers.
- Process metrics.

Add process metrics registry & collection task for Stream & Operator.

Updated Stream component to spawn the process metrics sampler routine.

Updated Operator component to spawn the process metrics sampler routine.
Also exposing the /metrics & /health endpoints on a standard HTTP (no TLS)
server. This reduces a bit of the overhead in configuring monitoring and
the like.

Hadron Core has been updated to properly compile and expose process
metrics gathering when running on a linux OS.