I am getting a 503 error while using Azure OpenAI very frequently.

Arpan Malik 5

Model is GPT4 and region is South India. I am using a standard deployment with 10k tokens /min. Trying to use the llm on a pyspark dataframe using a langchain pyspark agent.
The error looks like this:
InternalServerError: Error code: 503 - {'error': {'code': 'InternalServerError', 'message': 'The service is temporarily unable to process your request. Please try again later.'}}

dupammi 8,535 Reputation points Microsoft Vendor

2024-06-21T09:37:58.8333333+00:00

Hi @Arpan Malik

Thank you for your question.

The 503 error you are encountering while using Azure OpenAI with a GPT4 model in the South India region indicates that the service is temporarily unable to process your request. This could be due to a variety of reasons, such as a resource constraint. I checked the Azure status page at https://status.azure.com/en-us/status and didn't find any know issues. Also check model-summary-table-and-region-availability

I recommended that you check your quota and limit in Azure OpenAI Studio to ensure that you have not exceeded your quota. If everything looks fine, you can try again later as the error message suggests. If the issue persists, you can contact Azure support case through Azure portal

Thank you.
Arpan Malik 5 Reputation points

2024-06-21T12:55:24.95+00:00

Hi @dupammi

I checked the quota limit and its 300k. I am using only 100k.
Also the model is available in South India.

This error is coming suddenly while using the API and then works again after some time. As this API is directly used in a UI, I can't afford to encounter this every now and then.
Bruno Goncalves Vaz (P) 20 Reputation points

2024-07-03T09:31:34.9133333+00:00

I'm having the same problem with the Whisper model through Azure OpenAI.

Share via

I am getting a 503 error while using Azure OpenAI very frequently.

Your answer