Add metrics for chain fork work and lengths

[teor2345]

Motivation

The ZIP editors asked for the following Zebra metrics:

• histogram of chain work for all recent chain forks
• histogram of tip heights for all recent chain forks
• histogram of non-finalized lengths for all recent chain forks
• histogram of fork heights for all recent chain forks
• histogram of fork lengths (tip height - fork height)
• all the above metrics as a gauge for the best chain only

In #6235, we added APIs for these metrics, but we need to track them using prometheus in this ticket:

• fork heights
• fork lengths

We already implement these metrics

• number of recent forks
• some of the above metrics (for the best chain only)

zcashd is currently unable to track these metrics, because it doesn't seek out multiple chains in parallel.

[mpguerra]

When would they like these by? The earliest I would schedule these would be next sprint (21) and only if the release candidate work is complete

[daira]

This is for ZIP 315, to verify whether the recommended "trusted" and "untrusted" confirmation depths make sense and to provide supporting rationale. So not immediately urgent.

[teor2345]

This is blocked by PR #5257 merging, because it changes some of the same code as this ticket.

[mpguerra]

Hey team! Please add your planning poker estimate with Zenhub @arya2 @conradoplg @oxarbitrage @teor2345 @upbqdn

[mpguerra]

tentatively scheduling this for after the GBT work

[teor2345]

Some of these changes are already made in PR #6235, when I get time, I'll split them out into a separate PR.

[teor2345]

Some of these changes are already made in PR #6235, when I get time, I'll split them out into a separate PR.

I updated this ticket with the new metrics APIs that are part of #6235, and the extra things we need to do in this ticket.

[teor2345]

We probably want to delay this until after we've renamed some parts of the state in response to an audit finding.

[teor2345]

This might also depend on #7109, because at the moment we're storing the partial work of the fork, not the cumulative work of the entire chain.