Question

A company processes payment events using an Amazon SQS standard queue that triggers an AWS Lambda function. Occasionally, a small number of malformed messages cause the Lambda function to fail repeatedly. These messages are automatically retried, consuming Lambda concurrency and delaying processing of valid messages. The team wants to isolate messages that cannot be processed after several attempts so they can be inspected later, without blocking the main queue. What should the solutions architect recommend?

Accepted Answer

Configure a dead-letter queue on the source SQS queue with a maxReceiveCount, so messages exceeding the retry threshold are moved off the main queue.

Answer

Increase the visibility timeout on the SQS queue so failing messages become visible less frequently.

Answer

Switch the queue to an SQS FIFO queue to guarantee exactly-once processing and prevent the malformed messages from being retried.

Answer

Enable Lambda provisioned concurrency to ensure enough capacity exists to process both valid and malformed messages.