Is your feature request related to a problem? Please describe.
Yes. It is related to a possible problem with CloudWatch logs. If for some reason there is an outage or a bad configured CWL sink (like non existing log_group). The sink must NOT have DLQ configured. On such scenarios the current configuration is to retry up to 15 times (configurable), but after that number of retries the original message is acknowledged and never tried again which means that we are dropping data to the destination.

Describe the solution you'd like
We would like that, except on certain terminal exceptions that depends on the data (data is too big, or too old or too new), when DLQ is not configured, to retry for ever.

Describe alternatives you've considered (Optional)
Increase the number of retries from 15 to a larger number. But those would means that there is no visibility (no metrics) about those exceptions happening.