Question

A company runs several Lambda functions in the same AWS account. One function, invoked by an SQS queue, can scale up rapidly during traffic bursts and overwhelm a fragile downstream legacy API that supports only 50 concurrent connections. During bursts, other unrelated Lambda functions in the account also start throttling because the account-level concurrency limit is being consumed. What is the MOST effective way to cap this one function's scaling while ensuring the other functions retain concurrency?

Accepted Answer

Configure reserved concurrency on the SQS-triggered function set to a value at or below the downstream API's limit.

Answer

Enable provisioned concurrency on the SQS-triggered function to pre-warm 50 execution environments.

Answer

Request an increase to the account-level concurrency limit through AWS Support.

Answer

Add a dead-letter queue to the SQS-triggered function to capture excess invocations.