Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
Azure Document Intelligence is updated continuously, and service changes can affect feature behavior and capability limits. This article tracks known issues in Azure Document Intelligence, describes their impact, and provides mitigation or resolution guidance. Before you submit a support request, review the list to determine whether your issue is a known condition and to identify the recommended remediation.
Service-level outages and notifications
Azure status and Azure Service Health publish service-level outages and regional disruptions. Use these services to validate current impact, track incident lifecycle updates, and configure proactive notifications for your subscriptions and resources.
- Service-level outages: Azure status page.
- Outage notifications and alerts: Azure Service Health Portal.
Current known issues
The following table lists active, service-confirmed issues in Azure Document Intelligence. Each entry includes the issue category, observed behavior, documented workaround, and initial publish date.
| Issue ID | Category | Title | Description | Workaround | Publication date |
|---|---|---|---|---|---|
| 1001 | File types | Content Is Not Supported Error With Excel Files |
Excel files are documented as a supported input type. Requests can fail with Content isn't supported when file content exceeds an internal character limit. This condition usually occurs with large or densely populated spreadsheets. | Excel files larger than ~8,000,000 characters are rejected. Option A: Split large Excel files into smaller files and submit them separately. Option B: Switch to Azure Content Understanding, which uses improved parser logic for these files. |
April 2026 |
| 1002 | Encryption | ServiceUnavailable error when using CMK |
Operations that rely on Customer‑Managed Keys (CMK) can intermittently fail with a service unavailable error. This failure occurs when the storage service uses stale cached identity or key material. | Option A: Refresh the CMK cache. Disable and re-enable CMK, assign a new key, or associate a new managed identity.Option B: Switch to Microsoft-Managed Keys (MMK): Navigate to your Document Intelligence resource in the Azure portal. Resource Management → Encryption -> Microsoft Managed Keys. |
April 2026 |
| 1003 | Model management | Model Exists Error When Model Copy |
Copying a custom model can fail with a Model Exists error. This error occurs when a broken or partially created model remains on the target resource. Customers can't resolve this condition themselves. | No customer-side workaround is available. Submit a support request so the product group can remove the broken model from the backend. |
April 2026 |
| 1004 | Model training | HTTP 500 – InternalServerError during model training (Neural or Template, Preview/GA APIs) |
Model training can fail with HTTP 500 / InternalServerError. This failure typically occurs when managed identity isn't configured correctly on the resource, especially in newer training flows. | Enable Managed Identity on your Document Intelligence resource. Verify that the identity has the required permissions. Follow the official managed-identity configuration guidance. |
April 2026 |
| 1005 | Model training | Model training stuck in *Not started* status |
If a model training job fails before execution, the failed job may remain visible in the model list with status Not started. This state results from backend lifecycle handling, not customer configuration. | The backend automatically clears failed training jobs after approximately seven (7) days. If earlier removal is required, submit an Azure portal support request. CSS or PG engineering can execute backend deletion of the orphaned training record. |
April 2026 |