We have made a handful of improvements to our rate limits to help maintain a high-quality API and support optimal performance. Some of these updates include:
Standardized error codes: Rate limit errors will now return a 429 HTTP error code to help standardize responses and simplify error handling.
Retry-After header: Errors will now include the Retry-After header to indicate how long you need to wait before making another request.
New rate limits: We’ve introduced new limits, including per-minute request limits and a concurrency limit. These changes help reduce the frequency of complexity limits, giving you more control over your API usage and providing a smoother developer experience.
Daily limits: Daily request limits will be enforced for basic and standard accounts starting next week. Limits for pro and enterprise accounts will be gradually rolled out next year.
For more information, check out our rate limits documentation.
Daily limits: Daily request limits will be enforced for basic and standard accounts starting December 10th, 2024. Limits for pro and enterprise accounts will be gradually rolled out next year.
(Rate limit improvements)
I have serious concerns regarding this as our automations are runnning throughout the whole day for various clients. What will the limit on those be? For enterprise and pro accounts? Will this be connected to the account_id or the API key?
Hi Daniel.
Thank you, but I am not sure how the recommended usage for enterprise accounts
(in our case 400+ users) is 25k requests PER DAY.
For example:
If each of our users takes a look at their time logging board (hundreds of items) with pagination it gives about 100 requests per user, sometimes more, depends on the number of items in the boards. 100 * 400 = 40k. And that’s only if they take a look at the app once. But other than that we also have updating items, creating items, nightly processes that are reading and updating items, also in hundreds of items.
I want to provide some clarity on the recent announcement regarding the introduction of a daily API limit. I want to assure you that our primary objective is to ensure a smooth and seamless transition for all account, with no immediate impact on your current usage.
This new policy applies only to personal account API usage and not to marketplace app developers. For the time being, API usage by marketplace apps will not count towards this limit. Even when the limit is reached (and in cases where it is enforced), marketplace apps will remain unaffected.
For those accounts that will be impacted by this change, we have planned a two-step approach to ensure a smooth transition:
We will provide a new observability dashboard for API usage, allowing you to monitor your usage and understand where the API is being utilized.
If we anticipate any impact on your account due to this change, we will engage in personal communication and provide ample time for adjustments.
For our Pro and Enterprise tier users, this will be a soft limit for the foreseeable future to prevent any disruption to your workflows. If your account requires a higher API limit we will proactively adjust it based on the size and needs of the account.
Our ultimate goal remains unchanged - to continue providing an exceptional API experience that scales with your needs, and in the future is also appropriate for the tier and size of the account.
We appreciate your understanding and cooperation as we implement this new policy. As always, our technical support team is available to answer any questions or concerns you may have.
Daily limits do not apply to marketplace apps. This is good.
Question about the minute limit and concurrency limits. Are these per app / user token per account or are they simply per account regardless of the source? (meaning different apps are competing with each other for these resources)
Thanks. I think the rate-limits documentation page needs improvement to clarify the scope of each limits.
There is even an interpretation that some limits aren’t per account. And an app can only make 500 mutations per minute globally to the platform, or x complexity per minute.
Complexity limit implies per user.
Daily limit implies per account.
Minute limit could be interpreted as account, app, or user.
This isn’t new to these recent changes. The complexity limits often raised the same questions - esp. the distinction between app tokens and personal tokens (I think personal tokens on an account actually all share the same complexity budget - do they also share the minute limit and concurrency limits?
