If I have an Oban worker which uploads to s3, would rely on the exaws s3 client built in retries or would you rely on the Oban worker to retry when there is an s3 upload failure?
I suspect the best approach is to disable exaws retries and retry at the Oban worker level instead because the exaws retries probably will not finish retrying if there is deployment which shuts down the node in between exaws retries.