Push stats and metrics to daemon service

Metrics in current FDB are pulled by metrics consumer.

If each FDB process can push stats, metrics and incidents to some HTTPS or gRPC endpoint, standard metrics tools can consume them easier.

This can help operation tools (say instance lifecycle management tools) to automatically increase/decrease the resource for a component based on the metrics it receives.
For example, when a storage server's memory utilization is high, the server pushes an incident to the operation tool. The tool can start an instance with larger memory and move the storage server there.

[This idea was proposed and discussed in FDB Offset 2019.]

The issue names no files, tests, or entry points. Start by locating the current metrics consumer and how FDB processes expose stats, then review the Offset 2019 discussion for requirements. Done would require an agreed HTTPS or gRPC reporting design and its implementation, including stats, metrics, and incidents.

Written by the indexing model from the issue text.