I apologize if this question has been asked before.
Rate limits are at 10/second. I understand this. But is this 10/second/subscription, 10/second/application, or 10/second/environment (per app). Thank you for your time!
Per Subscription, see doc here.
Interesting. So this means each subscription does, in fact, have a limit of 864,000 per day. That means if a partner has 32 customers (environments), even through the partner has 100k calls per connection, the partner doesn't have 3.2 million calls per day, but only 864K. Sorry for pushing for clarity, but the design for our integration needs this detailed information. Thanks!
not sure if you are understanding this wrong, please read the doc in more details.
rate of 10 api calls per second is max rate in concurrency. there is a quota of 25K api call in total in a 24hr period for customer. it's less (10K i believe) if you are not a customer, or more than 25K if you pay for it
not sure what kind of integration you are building that requires this much calls, are you reading 1 record (constituent or gift) at a time?
I data sync into my data warehouse every 4 hours for 30+ tables (constituent, gift, opportunity, constituent code, constituent custom field, etc) and does not need to use this much calls.
@Thomas Royal - I'll have Stephen O'Neall reach out from the partner team to help clarify this for the Feathr team.
Thank's y'all. You're amazing.
This is a great question.
Erik mentioned the partner team would clarify - what was that clarification?
We are still waiting.
Hey @Thomas Royal and @Eric Eskildsen,
The quota and rate limits are applied on a per connection basis. A connection is between an app and an environment. So if there are two apps connected to the same environment, each app→environment connection will have its own quota and rate limit.
Hope that helps.
Thanks @Ben Wong !
Let me check my understanding here….I'll go back to @Thomas Royal 's example. I liked the numbers for reasoning about this.
A partner has 32 customers. Each gets their own quota and rate limit.
The partner makes parallel API calls. Let's say they need to max that out for whatever reason.
The partner could make up to 320 API calls per second across all customers. In other words, they could make 10 calls per second for each of their 32 customers, at the same time.
Their max API calls per day per customer would be 100K. So, across all customers, their systems could make 3.2 million daily API calls.
Is that right?
@Eric Eskildsen That sounds right in theory. Not all endpoints are equal in terms of response time so in practice, you can only make 10 calls per second if the endpoints can respond fast enough. Some endpoints are more resource intensive than others, so a lot of this depends on what APIs you are calling.
