Move backoff retry logic to retry requester
What does this MR do?
Move the backoff logic to retry the requester
- Moving all the retry/backoff logic in one place is required to introduce the retry metrics.
- Improvements have been made to backoff, instead of backing off on the blanket status codes, client will only backoff and retry on the status code for which retry makes sense.
- Rate limit requester has been renamed to retry requester, which is more appropriate.
Why was this MR needed?
Required to add retry metrics.
What's the best way to test this MR?
go test -race ./network/...
What are the relevant issue numbers?
Edited by Love Bhardwaj