Add robust Groq rate limiting

[talhasarit]

Summary

• add a SQLite-backed Groq transcription limiter that enforces per-model request and audio-duration quotas before uploads are sent
• wire the limiter into the transcription fallback path, track reservation outcomes, and add deterministic tests for single-worker and concurrent worker scenarios
• add a read-only /jobs/groq/rate-limit endpoint plus docs so current limiter usage and remaining headroom are easy to inspect

Test plan

• pytest

Made with Cursor

[cursor]

You have run out of free Bugbot PR reviews for this billing cycle. This will reset on April 7.

To receive reviews on all of your PRs, visit the Cursor dashboard to activate Pro and start your 14-day free trial.