Given that this seems to be exposing a legitimate issue with the underlying service, do we still want to try to work around it with retries in tooz?
Given that this seems to be exposing a legitimate issue with the underlying service, do we still want to try to work around it with retries in tooz?