Hi Team, We use MSQ to ingest data from S3 into Druid and some of the tasks are failing with a "503 ServiceUnavailable" error (request rate limit breach). I came across this old issue which is supposed to be fixed in a subsequent PR. Issue: https://github.com/apache/druid/issues/10289 PR: https://github.com/apache/druid/pull/10761 But I don't see any retries/backoff happening in the task logs. Do we need to configure the retry/backoff explicitly? I couldn't find any params in the S3 extension to do this. Druid Version: 24.0.2 Regards, AR.