Lambda function has started to throttle requests

Why do I see this?

One of your Lambda functions has throttled request(s) within the last one hour.

What does this mean?

In AWS Lambda, concurrency is the number of requests that your function is serving at a given time. By default, AWS Lambda has a soft limit of 1000 concurrent executions per region. This limit applies to all of your functions in the same region.

From this pool of 1000 concurrent executions, you can set aside a reserved number of executions for any particular function using the concurrency configuration option available in the AWS console.

If your function exceeds the configured function-level concurrency limit, the AWS Lambda service will start throttling that function by rejecting any further requests.

This event notifies you when one of your Lambda functions has throttled request(s) within the last one hour.

How do I fix "Lambda function has started to throttle requests"?

If you were notified of this event for any of your Lambda functions, it means that your function has reached the maximum number of concurrent executions that was configured for it. You will need to re-adjust the configured concurrency limit.

You can adjust the function-level concurrency of your function using the concurrency configuration option in the AWS console. If you are hitting the AWS region limit of 1000 concurrent executions, you can request a concurrency limit increase by opening a quota increase case in the Service Quota dashboard in the AWS console.

To determine the required concurrency limit, you can multiply your function's average runtime duration in seconds by the average number of requests per second.