http: request queue and timeouts

[mastercactapus]

• Identified the issue which this PR solves.
• Read the CONTRIBUTING document.
• Code builds clean without any errors or warnings.
• Added appropriate tests for any new functionality.
• All new and existing tests passed.
• Added comments in the code, where necessary.
• Ran make check to catch common errors. Fixed any that came up.

Description:
This PR introduces request queueing to the concurrency limiter.

The concurrency limiter was designed to limit the number of simultaneous integration keys, service, and/or user-requests that are processed concurrently per-instance.

The present-day implementation limits concurrency, effectively isolating busy consumers, but all "pending" requests are then processed in random order. This results in latency spikes under load (per individual key) and 502 responses due to "backend timeout".

This change introduces a FIFO queue for all pending requests instead of the random group.

• For requests beyond the maximum (250 per key), they are immediately returned with status code 429 (too many requests)
• Requests stuck in the queue for longer than 25s are also returned with 429

Rejected requests are still scoped per integration key/service/user as before but will now allow a distinction between actual hung requests and throttling.

Describe any introduced API changes:
Throttled requests will now return status code 429 instead of timing out and returning 502.

[m17ch]

LGTM