We get an Internal Server Error on Azure Cognitive services

Johan Coens 5 Reputation points
2026-01-15T16:22:15.95+00:00

Since this morning we run into InternalServerError on Azure AI Vision Text searches. We use model 2023-04-15 for +1 year now, but stopped working this morning. The legacy model version 2022-04-11 does work but is not suitable for us.

We see no outage reported on Azure status, but are unable to resolve it ourselves.

Azure AI services
Azure AI services
A group of Azure services, SDKs, and APIs designed to make apps more intelligent, engaging, and discoverable.
{count} vote

3 answers

Sort by: Most helpful
  1. Deleted

    This answer has been deleted due to a violation of our Code of Conduct. The answer was manually reported or identified through automated detection before action was taken. Please refer to our Code of Conduct for more information.


    Comments have been turned off. Learn more

  2. Deleted

    This answer has been deleted due to a violation of our Code of Conduct. The answer was manually reported or identified through automated detection before action was taken. Please refer to our Code of Conduct for more information.


    Comments have been turned off. Learn more

  3. Sina Salam 27,786 Reputation points Volunteer Moderator
    2026-01-29T10:14:30.56+00:00

    Hello Johan Coens,

    Welcome to the Microsoft Q&A and thank you for posting your questions here.

    I understand that you are getting an Internal Server Error on Azure Cognitive services.

    To resolve that, execute the below steps in order:

    1. Switch your Vision endpoint to a healthy region (e.g., West Europe/East US). - https://learn.microsoft.com/en-us/answers/questions/2115918/quickstart-optical-character-recognition-%28ocr%29-int
    2. Keep Read v3.2 with model-version=2022-04-30 (documented GA) on /vision/v3.2/read/analyze. Do not use 2023-02-01-preview for Read v3.2—it’s not in the official list. - https://learn.microsoft.com/en-us/azure/ai-services/computer-vision/how-to/call-read-api
    3. Add 3 retries (exp backoff), plus a circuit breaker to stay on the secondary region while the primary is degraded. - https://learn.microsoft.com/en-us/answers/questions/5747009/azure-openai-doesnt-work-region-sweden-central
    4. Log request IDs (x-ms-request-id, apim-request-id) and open a Severity A ticket with exact timestamps and comparison data (legacy/other region). - https://learn.microsoft.com/en-us/answers/questions/1497324/internal-server-error-when-calling-cognitive-servi, https://learn.microsoft.com/en-us/answers/questions/5742974/we-are-encountering-an-internalservererror-%28http-5
    5. Enable Service/Resource Health alerts to get real‑time notifications and model lifecycle notices for proactive changes. - https://learn.microsoft.com/en-us/azure/ai-foundry/openai/concepts/model-retirements?view=foundry-classic

    This plan above will restore service within the hour and it align strictly with current Microsoft documentation and incident reports.

    I hope this is helpful! Do not hesitate to let me know if you have any other questions or clarifications.


    Please don't forget to close up the thread here by upvoting and accept it as an answer if it is helpful.

    0 comments No 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.