API calls per minute
Are all REST requests counted towards that limit?
I know we can cache reads, but it really limits what we can create (within reasonable cost).
Even at 200$/mo we can do max 41 requests per second, even a 5$/mo VPS can do that and more.
What happens when we reach the limit and how do we know the limit has been reached?
I didn't think of the Amazon API Gateway, which is required for every Lambda (API request) to run.
So 41req/s would actually cost you 122$/mo (+ traffic) management overhead excluded, so now I see why you charge for it and why there's a limit.
Problem is, such loads are rarely sustained - would rather see a request-count limit, rather than a per-minute limit.
Joseph Benguira last edited by
Indeed hosting on AWS Serverless infra cost more but is a lot more reliable than a $5 vps ... and is fully managed by AWS team directly
FYI you can do "Burst" and use up to 5 minutes of credits
so let say your subscription allows 1000 RPS, you are allowed to burst up to 5000 RPS!
When you are reaching the limits (over the burst) you will automatically receive an email offering you to upgrade to an higher plan.
while you are over the burst limit our API will return an error saying you are over quota
after the calculation period (5 minutes) your api will resume serving requests automatically
You really thought of everything with this!
I really appreciate the transparency too, other services charge several hundred percent for made up limits that don't really have any anchor in reality other than "because".
Daniel Mulroy last edited by
@Joseph-Benguira Can you double check your math?
You're saying "RPS" (Request Per Second), but saying you can use 5 minutes of credit and the example you gave is 1000 RPS -> 5000 RPS.
Did you mean 1000 RPM -> 5000 RPM?
Wassim last edited by