These 502's are continuously alerting via sentry. Would it be ok to either add retries for 502's back or allow a way to inject my own middleware to the client?
I am happy to make a PR for either. I don't want to just silence the 502's in case there is an actual problem I would want to be alerted.
Hi 👋
I am having the same issue as this previous issue: https://github.com/dashbitco/broadway_cloud_pub_sub/issues/55
The fixed seemed to be adding retries to Tesla.
Tesla and retrying was refactored away when the client changed to Finch: https://github.com/dashbitco/broadway_cloud_pub_sub/pull/66
These 502's are continuously alerting via sentry. Would it be ok to either add retries for 502's back or allow a way to inject my own middleware to the client?
I am happy to make a PR for either. I don't want to just silence the 502's in case there is an actual problem I would want to be alerted.