GitLab CE/EE timeouts should be the same for Sidekiq and Unicorn
Everyone can contribute. Help move this issue forward while earning points, leveling up and collecting rewards.
gRPC has a timeout setting for each RPC performed. This allows the client to give up on a request, signal the server it doesn't need the response anymore, and continue the execution. For Sidekiq and Unicorn there's different timeouts, mostly as with Unicorn we want to give up on the request before we kill the worker and Sidekiq is performing long running jobs that should only time out after a very long time.
Right now the sidekiq server doesn't have a timeout associated with a request, only if the RPC uses the default timeout. This has some surprising side effects, as noted by @smcgivern in: https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/20974#note_92294773
We should have the same behavior for both Sidekiq and Unicorn, if possible. Or include other timeout values for all 3 the methods, fast, default and slow timeout.