Exceptions Handling
GenAI Azure OpenAI Inference API - Troubleshooting
Should you encounter issues while interacting with the Azure OpenAI services via TrustNest APIM, this section provides guidance on identifying and resolving common problems.
1. Authentication Errors:
If you receive 401 Unauthorized
error, check the following:
- Ensure your
TrustNest-Apim-Subscription-Key
andAuthorization
token are correct. - Verify that your account has the necessary permissions to access the specified endpoints.
- Make sure your token has not expired. If it has, generate a new token by going to your profile and clicking on "Regenerate" in the Subscriptions section.
2. Invalid Request Errors:
Receiving a 400 Bad Request
error usually indicates a problem with the request's syntax or structure:
- Verify that your request adheres to the correct format as outlined in the endpoint documentation.
- Ensure all required parameters are included and are correctly formatted.
3. Rate Limit or Quota Exceeded Errors:
A 429 Too Many Requests
error suggests that you've hit the rate limit for the API:
- Review the rate limits in the Azure OpenAI Service quotas and limits documentation.
- Implement a backoff strategy in your code to respect the rate limits.
4. Internal Server Errors:
A 500 Internal Server Error
suggests a problem on the server-side:
- Verify that your JWT token is valid and has not expired. If it has, generate a new token.
- Retry the request after a brief wait.
- If the problem persists, check the Status Page for any ongoing incidents
5. Service Unavailable Errors:
A 503 Service Unavailable
error suggests the service is temporarily unavailable:
- Retry the request after a brief wait.
- If the problem persists, check the Status Page for any ongoing incidents.
6. Content Filter Errors:
If your request triggers a content filter error 400 Content Filter Error
, check the following:
- Review and modify the content of your request to comply with the Azure Content Safety & Filtering guidelines.
7. Debugging:
- Inspect the error message returned by the API as it often contains useful information about what went wrong.
- Check the API documentation for the endpoint you are trying to use to ensure your request is structured correctly.
8. Getting Help:
- Consult the OpenAI Error Handling Documentation for more detailed information on error codes.
- If you are unable to resolve the issue, consider reaching out to Azure OpenAI support or TrustNest support for assistance.