Add a prometheus metrics endpoint for monitoring

> [!NOTE]
> Migrated from [augurlabs/augur#3600](https://github.com/augurlabs/augur/issues/3600)
> Originally opened by `@MoralCode` on 2026-01-16

---

Originally part of #159 but split into a separate issue. Its mostly a list of ideas for things that could also make a good prometheus metrics endpoint:
- report aggregate stats about API keys such as total remaining API calls by rate limit group (the verification script could use this to make a summary (i.e. "you have X valid API keys loaded (and Y invalid ones). ZZZZ REST and WWWW GraphQL requests can be made across all of these keys. This indicates that your install has (high/medium/low/no) API calls left compared to the total exected capacity of all keys")
- report statistics about collection (this is hard for the grafana stack to do on its own since its kind of a 3d/4d query if you have variables for # of days to consider "recent", and whether you want to look at a snapshot of all things or a time series of one thing)
- count of repos collected in each phase today
- count of total configured repos
- count of ignored repos
- count of repos with failures by collection status
- ... this week/ two weeks, etc
- earliest, quartile, median, and latest dates of collection for each phase of collection

This would be a nice thing that would improve the grafana stack but also allow others to use it for their own monitoring too.