Feature Request

Docker Compose to start Granafa and Prometheus servers that monitor locally running nodes.

Summary

The Grafana dashboards in the qa-infra repository have some interesting graphs that we could use during local testing. We could also improve those dashboards easier if we have a locally running Grafana server.

Problem Definition

I have a hard time understanding the health of a node in some cases: how long does consensus take, what's the memory usage, etc. The Comet log file has a lot of information but we should leverage the Prometheus metrics during development better.

With this tool we'd get easier visibility into Prometheus metrics. It's an external tool, no Go code needs to change.

Proposal

• Docker Compose file that starts up a Grafana and a Prometheus container.
• They need to connect to each other.
• Prometheus needs to be able to connect to nodes running locally on the host machine.
• Pre-populated dashboards would be nice.