Question

A developer configures an AWS Lambda function with an event source mapping to a standard Amazon SQS queue. The function processes each message by writing to a legacy on-premises API that can handle only a limited number of concurrent connections. During load tests, the developer observes that Lambda scales up rapidly and overwhelms the API. The team wants to limit how many messages Lambda pulls and processes at once without dropping messages. Which combination of settings should the developer adjust?

Accepted Answer

Set a reserved concurrency limit on the Lambda function to cap the number of concurrent executions.

Answer

Increase the batch size on the event source mapping so fewer invocations are made.

Answer

Enable provisioned concurrency on the function to pre-warm all needed instances.

Answer

Reduce the SQS queue's visibility timeout so messages are retried faster.