Conduit: Prometheus metrics endpoint

[winder]

• Metrics:
  • There is no metrics endpoint.
  • Missing metrics: durations for each plugin, duration for processing, total round processing time, ...

We should avoid using echo for this. Since there is a single endpoint something simpler should be sufficient.

Acceptance Criteria

1. Create an HTTP server that hosts Prometheus metrics.
2. Add new pipeline configuration to enable metrics (yaml file, not CLI flag).
3. Hook it up to the current metrics that Indexer uses.
4. Add new metrics that are Conduit pipeline specific (i.e. duration each plugin runs for)
5. Provide an interfaces to plugins, so that they can store their own metrics.