Add adaptive request concurrency
What does this MR do?
Adjusts the request concurrency for jobs based on demand.
Why was this MR needed?
- Reduces the number of requests
- Reduces reservations, so that we don't keep idle instances for longer than needed
What's the best way to test this MR?
Test with runner with a request_concurrency of 1 and a request_concurrency of 25 and compare.
What are the relevant issue numbers?
Closes #37237 (closed)
Edited by Arran Walker