Question

A developer configures an AWS Lambda function with an SQS standard queue as its event source mapping. The function processes messages in batches of 10. Occasionally, one or two messages in a batch fail while the others succeed. Currently, when any message fails, the entire batch is returned to the queue and reprocessed, causing already-successful messages to be processed multiple times. What is the recommended way to ensure only the failed messages are retried?

Accepted Answer

Report partial batch failures by returning the failed message IDs in a batchItemFailures response and enabling ReportBatchItemFailures on the event source mapping.

Answer

Set the batch size to 1 so that each invocation processes a single message and failures never affect other messages.

Answer

Increase the visibility timeout on the SQS queue so successful messages are not reprocessed before the batch completes.

Answer

Configure a dead-letter queue on the Lambda function so failed messages are moved out of the batch automatically.