Box Content API detailed throttling logic

We are building an application using Box Content API to scan users' content. As multi-threading is used, we see a lot of "429 too many requests" errors. Unfortunately, there is not much detailed information about how API throttling is handled at backend, we just saw some posts talking about 50000 per day limit.

From our observation, however, it seems to be related to the number of API calls made within a given period of time instead, as well as number of users and apps. Could someone help to share some insight on this? for example, would using multiple users concurrently help the situation; would admin impersonate different users count as multiple users or must get token from multiple users from the beginning?

The detailed understanding would allow us to see how to align our application structure accordingly. Thanks in advance and any input would be greatly appreciated!

Brent

April 18, 2016 13:42

Thanks for your question! Public forums do not accurately represent the Box Content API rate limits. You are on the right path related to the limits being set based both on the rate of actions by a user. For specific questions related to the Box API rate limits, please submit a support case here.

scarceller

February 09, 2018 07:04

Does/can Box offer un-limitted throttling? or very high throughput rates for large Box customers? I'd like to get around 5000 or more SDK queries per minute, current testing stalls me at around 250/minute.

Does Box offer custom throttle rates? even if they cost money on a very large Production ready system to support 50,000 employees?

Thanks in advance.

Comments