jli

If we have this, I'm fairly sure we should also decouple bulkheads from circuits. At the moment (or at least as of the end of 2018), a bulkhead acquisition failure...

I think we should expose enough API to implement either, and then have the driver implement the actual listener(s) and/or polling workers. In fact, I think the API would be...

Though that would probably be a good thing, security-wise.

Referencing #38 so it links here.

PUT overwriting everything seems fine to me, we can add POST routes that incrementally update things.

If access control is the frontend's job here, it needs to understand what a full scrape is anyway, so the config flag for enabling it at all could live there....

If profiling has shown debug logging that's getting thrown out anyway is what's taking resources here, then some dumb `if debug { log debug message }` boilerplate might be the...

I’ve had good results from using this library at work 👍

This should probably be at least configurable for tracker operators who want to support those clients, which are relatively common in the wild.

Like you said, replication delay isn't an issue, so I think a ring or otherwise loosely connected topology would be perfect. Something like each tracker connecting to 2 others would...