Share via

Document Intelligence prebuilt-read with api-version=2024-11-30 stuck in "running" (West Europe)

Mr B 0 Reputation points
2025-12-05T17:05:02.62+00:00

Service: Azure AI Document Intelligence (prebuilt-read)

Region: West Europe

API version: 2024-11-30

Since today, the prebuilt-read model with api-version=2024-11-30 is not completing for me in West Europe.

Repro #1 – public test file

Test document (PNG with text):

https://www.syncfusion.com/blogs/wp-content/uploads/2023/10/Performing-OCR-on-a-rotated-page-of-a-PDF-document.png

Working case (older API):

POST https://westeurope.api.cognitive.microsoft.com/formrecognizer/documentModels/prebuilt-read:analyze?api-version=2023-07-31

Content-Type: application/json

Ocp-Apim-Subscription-Key: <redacted>

Body:

{ "urlSource": "https://www.syncfusion.com/blogs/wp-content/uploads/2023/10/Performing-OCR-on-a-rotated-page-of-a-PDF-document.png" }

This returns 202 Accepted and the corresponding analyzeResults call quickly returns:

{ "status": "succeeded", ... }.

Failing case (new API):

POST https://westeurope.api.cognitive.microsoft.com/documentintelligence/documentModels/prebuilt-read:analyze?api-version=2024-11-30

Content-Type: application/json

Ocp-Apim-Subscription-Key: <redacted>

Body:

{ "urlSource": "https://www.syncfusion.com/blogs/wp-content/uploads/2023/10/Performing-OCR-on-a-rotated-page-of-a-PDF-document.png" }

The POST returns 202 Accepted and an operation-location header.

However, repeated GET requests to that operation-location always return:

GET https://westeurope.api.cognitive.microsoft.com/documentintelligence/documentModels/prebuilt-read/analyzeResults/931f46a7-32a1-4cdb-af0c-8e556a3d9e39?api-version=2024-11-30

Ocp-Apim-Subscription-Key: <redacted>

Response (even after 20+ minutes):

{

"status": "running",

"createdDateTime": "2025-12-05T15:30:27Z",

"lastUpdatedDateTime": "2025-12-05T15:30:27Z"

}

The status never changes to "succeeded" or "failed".

My Blob Storage and SAS URLs work fine:

  • I can upload the document to my storage account.
  • I can download the same blob via SAS with HTTP 200 and correct content.

When I switch back to the older API:

POST https://westeurope.api.cognitive.microsoft.com/formrecognizer/documentModels/prebuilt-read:analyze?api-version=2023-07-31

with the same document, the operation completes quickly with status "succeeded".

Question:

Is there a known issue with api-version=2024-11-30 for prebuilt-read in West Europe that could cause operations to be stuck in "running" state?

Can someone from the product team check why operation

931f46a7-32a1-4cdb-af0c-8e556a3d9e39

never transitions to succeeded or failed?

Azure AI services
Azure AI services
A group of Azure services, SDKs, and APIs designed to make apps more intelligent, engaging, and discoverable.
0 comments

1 answer

Sort by: Most helpful
Most helpful Newest Oldest
  1. Q&A Assist
    2025-12-05T17:05:21.1733333+00:00
    AI generated content. This question contains an answer created with AI from Q&A Assist. Learn more

    There may be several reasons why the prebuilt-read model with API version 2024-11-30 is stuck in the "running" state in West Europe. Here are some considerations:

    1. Service Availability: Sometimes, specific regions may experience temporary issues or outages. It would be beneficial to check the Azure status page or any announcements related to service disruptions in the West Europe region.
    2. Model Performance: The new API version may have different performance characteristics compared to the older version. If the document being processed is complex or large, it might take longer to analyze.
    3. Request Limitations: Ensure that you are not hitting any request limits or quotas set by Azure for your subscription, which could affect processing times.
    4. Document Format: Although you mentioned that the document uploads and downloads correctly, ensure that the document format and content meet the requirements specified for the prebuilt-read model.
    5. Polling Frequency: When polling for the status of the operation, it is recommended to wait at least one second between requests to avoid overwhelming the service.

    If the issue persists, it may be beneficial to reach out to Azure support for further assistance, as they can provide more detailed insights into the operation ID and any underlying issues affecting the model's performance.

    References:

    0 comments

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.