Concurrent execution

AWS Lambda dynamically scales capacity in response to increased traffic. However, there's a limited number of an executed function's code at any given time. This number is called concurrent execution, and it's defined per AWS region. The default limit of concurrency is 1,000 per AWS region. So, what happens if your function crosses this defined threshold? Read on to find out.

Lambda throttling

Lambda applies throttling (rate limiting) to your function if the concurrent execution count is exceeding the limit. Hence, the remaining incoming requests won't invoke the function.

Note

The invoking client is responsible for retrying the failed requests due to throttling by implementing a back-off strategy based on the HTTP code returned (429 = too many requests). It's worth mentioning that Lambda functions can be configured to store unprocessed events, after a certain number of retries, to a queue called the dead letter queue.

Throttling might be useful in some cases, as the concurrent...