SQS Queue: maxRetries when error handling inside worker, or customize Message retention period

[politoedo]

Start from the Use-case

I am using a FIFO SQS that triggers a Lambda function.
The lambda just copy and paste between buckets. The sequential order of message processing is vital.
Inside my lambda there are couple of exception handling that send back the message to the queue, enforce not processing subsequent messages having same group id, and send an error.

Since the error handling is performed inside worker (as it should be? honest question), the maxRetries policy does not apply. So I am stuck with a message that tries to get consumed every timeout*6 for 4 days, whereas a much simpler (((timeout*6)+1)*N) would be enough (in my use case). In general, it would be nice to have either (i) maxRetries being applied on error handling inside worker; or (ii) being free to set a Message retention period from our configuration.

Example Config

No response

Implementation Idea

Either implement maxRetries when error handling is done inside worker, or allow customization of Message retention period