Content Understanding Analyzer not working due to internal server error

Question

I just deployed a new content understanding analyzer and now I always get the error message "InternalServerError An unexpected error occurred." with an apim request id "444f011b-52a2-4722-9f07-19b642822e0b". I tried using a predefined scheme and I tried to update my scheme but nothing seems to be working. I also tried a http request but there I'd get a 500 error code.

Just last week the service was down for a while too and I was wondering if it is the same right now as https://azure.status.microsoft/en-gb/status indicates it's not, but last time when the service did not work it was also not indicated there.

Is this service even stable and suitable for production use? Two noticeable long downtimes are not really that great to rely on.

Answer 1

Hi ,

Thanks for reaching out to Microsoft Q&A.

Root Cause Possibilities

1. Backend Processing Issue: A consistent 500 Internal Server Error from multiple access points (portal UI and direct HTTP calls) points to a backend bug or service outage, not a configuration error on your part.

Deployment Glitch or Schema Issue: Sometimes analyzers deployed with an invalid schema or incompatible document format cause internal failures instead of graceful errors.

1. Service Instability (unlisted outage): azure's public status page might show only broad regional issues. A specific service like Content Understanding (especially in preview/private offerings) may face outages not reflected there.

Recommendations

Raise a Support Ticket with Microsoft:

Share your APIM request ID: 444f011b-52a2-4722-9f07-19b642822e0b.

  Include the screenshot and details about the schema/documents.
  
     Request backend log analysis for root cause.
     

1. Use Azure Resource Health: Check Resource Health for your specific Cognitive Services instance, it may show issues not listed publicly.
2. Retry in Another Region: If possible, deploy a test instance of the analyzer in a different region (West US or North Europe) to check if this is region-specific.

---

Stability for Production Use?

Currently:

The tool seems unstable and not fully production-grade.

If you observed multiple downtimes, and now repeated internal server errors, it is not ready for mission-critical deployments without fallback or monitoring logic.

You may want to:

• Log and monitor failures via APIM request IDs.
• Build retry logic and alerts.
• Keep a backup workflow (fallback to Form Recognizer or human review pipeline) if critical.

Please 'Upvote'(Thumbs-up) and 'Accept' as answer if the reply was helpful. This will be benefitting other community members who face the same issue.