Description of the new feature/enhancement

Given a large number of PRs in a short period of time, the pipelines can become somewhat unresponsive. When the back-end infrastructure starts responding with 503 and specifying Retry-After, things can get backed up. This can result in a cascade failure where it takes a long time to catch back up.

I'd like to have a better mechanism to queue and schedule validation runs to keep throughput optimized and avoid cascade failures where subsequent attempt to trigger validation runs don't continue to overload back-end services.

Proposed technical implementation details (optional)

No response