Application and Operational Metrics

[tmcgilchrist]

ocaml-ci should report on application and operational level metrics via prometheus/grafana.

• Production dashboard in grafana for https://ocaml.ci.dev
• Setup ocluster connection to export stats (it already supports prometheus)
• Add stat exports to ocaml-ci solver submissions
• Setup ocurrent to export stats (it already supports prometheus)
• Add ocurrent support for exporting Node creation/update/destroy metrics (as a proxy for graph size)
• GitLab pipeline reporting web hook events and rate limiting from GitLab.

Non-exhaustive list of tasks to cover the monitoring of ocaml-ci application.

[tmcgilchrist]

Provide a reason code for Cancelled jobs so we can tell the difference between Cancelled jobs that aren't required anymore and Cancelled because of timeouts.