Power BI ExportToFile API – Concurrent Request Throttling (429 Errors) and Request for Workaround

[poojagh]

Description:

We are encountering throttling issues while using the Power BI exportToFile API, despite adhering to the documented limitations.

According to the documentation:

• The exportToFile API supports up to 500 concurrent requests per capacity.

• To avoid Too Many Requests (429) errors, it's advised to distribute load over time or across capacities.

• Only five pages of a report are processed concurrently during export. For example, a report with 50 pages is processed in 10 sequential intervals.

• For performance, only a few export jobs should be executed in parallel at any given time.

Issue Observed:

• One of our team reported encountering 429 errors after only 20–25 concurrent export requests.

• In our internal load testing, we observed these errors after approximately 45 concurrent requests, which is significantly lower than the documented 50 (or 500 per capacity) limit.

• This behavior appears inconsistent with published limits and is impacting our ability to scale export operations.

Request for Support:

We would like the Power BI product team to:

1. Confirm whether any undocumented throttling rules or dynamic limits (e.g., tenant, workspace, or report complexity) might explain this early throttling.

2. Clarify if the 500 concurrent request limit per capacity is a hard cap or if there are other operational constraints that reduce the effective limit.

3. Provide guidance or recommendations on how to effectively manage and scale export workloads using this API.

Please let us know if more information is needed from our side. We would appreciate any help or alternatives the Power BI product team can provide to ensure consistent and scalable exports.

Thank you.

[v-ssriganesh]

Hi @poojagh,

Thank you for posting your query in the Microsoft Fabric Community Forum, and thanks to @AnkitKukreja for sharing valuable insights.

Could you please confirm if your query has been resolved by the provided solution? If so, please mark it as the solution. This will help other community members solve similar problems faster.

Thank you.

[poojagh]

Hi @v-ssriganesh ,
The query is still open and would like more insights on how this can be handled from a product point of view?
Any help here would be greatly apprecited!

[v-ssriganesh]

Hello @poojagh,
Thank you for the update.

Given that you're still experiencing throttling issues even with fewer concurrent ExportToFile API requests than the documented limits, it's understandable that you'd want clarity from a product behavior standpoint.

• The documented limit of 500 concurrent exports per capacity is a maximum theoretical threshold, assuming minimal capacity usage, lightweight reports, and no additional user or refresh loads.
• The Power BI service dynamically adjusts throttling thresholds based on Capacity load (CPU/memory), Report complexity (DAX, visuals, pages), Backend resource availability and Platform stability

This means 429 errors may be triggered earlier than expected, particularly during:

• Sudden bursts of concurrent exports
• Times of peak usage on the capacity
• Heavy or inefficient report exports (e.g., long DAX chains, DirectQuery, high page count)

To properly understand why you're encountering throttling below expected levels, we recommend opening a Microsoft support ticket for further investigation. You can explain all the troubleshooting steps you have taken to help them better understand the issue.

You can create a Microsoft support ticket with the help of the link below:
https://learn.microsoft.com/en-us/power-bi/support/create-support-ticket

If this information is helpful, please “Accept as solution” and give a "kudos" to assist other community members in resolving similar issues more efficiently.
Thank you.

[v-ssriganesh]

Hi @poojagh,

Could you please confirm if the issue has been resolved after raising a support case? If a solution has been found, it would be greatly appreciated if you could share your insights with the community. This would be helpful for other members who may encounter similar issues.

Thank you for your understanding and assistance.

[poojagh]

Hi @v-ssriganesh ,

The issue is still open

[v-ssriganesh]

Hello @poojagh,
We appreciate your patience and thank you for the update on the issue. A dedicated support team will be assigned to handle the break-fix scenarios efficiently. Please wait until the issue is resolved. Once resolved, kindly share the insights and mark them as "Accept as Solution" to assist others with similar issues.

Thank you.

[v-ssriganesh]

Hello @poojagh,

Just checking in to see if your issue has been resolved after raising the support case. If you’ve found a solution or any workaround, it would be wonderful if you could share it here. Your input could be valuable to others facing the same challenge.

Thank you for your time and contribution

[v-ssriganesh]

Hello @poojagh,

As per Microsoft community guidelines, we will now proceed to close this thread to keep discussions focused and manageable. If you still need assistance, you're welcome to start a new thread in the community at any time.

We appreciate your understanding and participation.
Thank you.

[AnkitKukreja]

Hi! @poojagh 

• The 500 concurrent requests per capacity is a theoretical upper bound under ideal conditions.
• In practice, effective limits can be much lower due to:
  • Report complexity (visuals, data volume, DAX calculations)
  • Capacity load (other workloads running)
  • Workspace or tenant-level constraints 

Export Power BI embedded analytics paginated reports API - Power BI | Microsoft Learn

• Microsoft applies dynamic throttling based on real-time capacity usage and system health.
• Even if you're under the 500-request limit, you may be throttled if:
  • Your capacity is under heavy load.
  • You're using Premium Per User (PPU), which has stricter limits (e.g., 1 export every 5 minutes per user)

Solved: Power BI Export to File API Rate Limitation - Microsoft Fabric Community

I would advice you to raise a support ticket in case you need anything else on this.