Add metrics to operator

[v9n]

Fixed ENG-844

Update the endpoint that generate the prometheus metric.

Convert all the metrics to vector to supports:

• operator address: This will allow us to later scrape and group by operator address.
• version: the version. Useful to find operator that has not update. Also being used on our telemetry page

(Version is just a number for demo purpose on this screenshot to show how it's appear)

Metrics generator:

• ping check duration|total, worker looper count
• Add "task received" metrics. Right now this isn't connected with real tasks yet, but the mechanism allow us to generate metrics, and therefor can update our dashboard.
• Add retrying in the sync loop so if the grpc stream is broken, we can re-connect to fetch new tasks

Change to connection pool:

• Record versions and metric port. We need the metrics port to later on connect the prometheus endpoint to our server to scrape all operator metrics, group by operator address.

[chrisli30]

Looks good!