Rate limiting documentation essentially states that rate limits will apply for API tokens
https://developer.atlassian.com/cloud/jira/platform/rate-limiting/
However, what's not clear is what exactly constitutes a bucket from the perspective of a token or credential. If there are multiple tokens for an account, there is ambiguity
- Is it one bucket per account, meaning that if the allocation is 400 per bucket, an account with 3 API tokens gets a bucket of 400 to share between all credentials?
- Is it one bucket per credentials, meaning that if the allocation is 400 an account with 3 API tokens gets a bucker of 400 for each credential?
That has significant implications for us for service accounts. We connect loads of tools to Jira.
Have you found the answer already? It seems that the bucket is associated with the site, not the account or the individual token.
When you have three tokens on a single account or spread across three accounts, they all draw from the same per-endpoint budget for that tenant once they visit the same site. This is particularly important for service accounts because splitting traffic among multiple tokens doesn't mean separate allowances; they still share one overall quota per endpoint on that instance.
I believe Atlassian confirmed this directly in a community thread when a customer asked a similar question about using separate accounts and tokens to isolate integrations: https://community.atlassian.com/forums/Jira-articles/Reminder-Please-ensure-your-Apps-comply-with-Jira-Cloud-Burst/ba-p/3150831
Cheers, Martin
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.