Amazon apigateway and lambda limits recommendation

Hi,

Is there someone who has deployed an api for a mobile app with 10K or similar users?
I would like to know if amazon default limits are reasonsable or they should be increased by contacting amazon support.
I know every case is different but I’m interesting in knowing your experience in order to extrapolate it into my case (it is photo stock app and also similar to instagram which is being developed currently)

Thank you in advance.

Amazon lambda limits (requests per second):
Throttle limits per account 1000 request per second (rps) with a burst limit of 2000 rps.
Amazon apigateway limit
Concurrent executions 100

1 Like

As you said, every use case is different. To make this somewhat easy to respond to, we would need to know if the 10k users is it per second, per hour, per day or per month.

Its also worth noting, that all of those limitations for lambda can be changed by sending an email or calling up AWS and having them increase your limits. They ask you a few reasonable questions, then increase it and you are now immediately scaled.

Just to update this question with the revised AWS Lambda / API gateway concurrent limits:

AWS lambda is 1000 concurrent executions per region (thats across all functions) by standard, with a burst increase to 3000 (main regions) & 1000 (new / smaller regions). Details here: http://docs.aws.amazon.com/lambda/latest/dg/concurrent-executions.html

AWS API gateway is 1000 rps (requests per sec), I can’t see any burst provisioning for API gateway in the docs, ref: http://docs.aws.amazon.com/apigateway/latest/developerguide/limits.html#api-gateway-limits

So for 10K users you should be ok, as long as you don’t have them all hitting your app at the same time. Best bet is to monitor your requests per sec average and use that to guide your provisioning (monitoring is your friend here :slight_smile:)

2 Likes