Handle more than once delivery in async workers

We have an issue where the async workers may fail in processing any given message once, but then succeed a second time. In these cases, we should do the simple thing and push to the user (or some retry logic) to create a secondary run. 