Service principal profile throttle limits

[Dan2]

When using a single service principal to access the powerbi API's there is a 120 request limit per minute. So for example if we create two service principal profiles as per example below, would this allow us to receive 120 requests per minute per service principal profile. Or does the 120 request per minute throttle apply to all of them together.

Example: 

Service principal master = 120 requests per minute

    - Service principal profile A = 120 requests per minute

    - Service principal profile B = 120 requests per minute

[v-tsaipranay]

Hi @Dan2 ,

Currently, there is no publicly available Microsoft documentation that details throttle limits for this scenario. As mentioned earlier, I suggest opening a support case with Microsoft so that a support engineer can review your environment and offer guidance specific to your situation.

Thank you.

[Dan2]

The problem is there is no solid documentation on the throttle limits. There is a dataset limit documented at 120 but this appears to apply to datasets only. The other powerbi api's, I have seen no documentation on the limits. The url I have found states there is a retry header that can be used as per url below. But would like to know a solid documented answer on what the limits are. Based upon my testing I get a 429 error from the powerbi api, if I exceed 60 calls per minute.

https://learn.microsoft.com/en-us/rest/api/power-bi/

[v-tsaipranay]

Hi @Dan2 ,

Could you please confirm if the issue has been resolved through the support ticket with Microsoft? If the issue has been resolved, we kindly request you to share the resolution or key insights here to help others in the community.

Please feel free to contact us if you have any further questions.

Thank you.

[v-tsaipranay]

Hi @Dan2 ,

To address your observation, the 120 requests-per-minute limit is enforced at the service principal level, not at the individual profile level. This means that all API calls made using any of the associated profiles, including the master, are collectively throttled under a single quota. Service principal profiles are designed primarily for organizational and tenant-level isolation, not for increasing or bypassing API throughput limits.

Regarding the 60 requests-per-minute cap you're observing per profile, even when tested in isolation this is not documented as a standard behavior and may be influenced by internal dynamic throttling mechanisms, tenant-level constraints, or load conditions specific to your environment.

To get a definitive answer about the throttling behavior you’re experiencing, particularly the 60 RPM limit, we suggest opening a Microsoft support case using the link below: How to create a Fabric and Power BI Support ticket - Power BI | Microsoft Learn

This will enable the Microsoft team to review backend telemetry for your tenant and provide an official explanation.

Hope this helps. Please reach out for further assistance.

Thank you.

[v-tsaipranay]

Hi @Dan2 ,

I wanted to follow up on our previous suggestions. We would like to hear back from you to ensure we can assist you further.

Thank you.

[Dan2]

Thanks for the answer. I wrote a test program that would use a master account and 3 profile accounts to hit the powerbi api's simutaneously. From the test results it looks like we get throttled at 60 requests per minute per account. Meaning each acccount/profile was allowed 60 requests per minute. When only running the master account or just a profile account it still had 60 per minute. So it appears tonot be ccumulative but also not the advertised 120 requests per minute. I just haven't found any concrete information out there that would definitively let me know how this is handled on the powerbi backend.

[v-tsaipranay]

Hi @Dan2 ,

Thank you for reaching out to the Microsoft fabric community forum.

To clarify, the 120 requests per minute limit applies to the entire service principal, not to each individual profile. This means that all requests made using any profile under the same service principal including the master, Profile A, Profile B, and so on are counted collectively within the same 120-request/minute limit.

If you’re looking to scale beyond this limit, we recommend using multiple distinct service principals or optimizing your request patterns to reduce frequency and improve efficiency.

Thank you.