Arize AI Community Icon

Home
Events

Arize AX Releases
Arize AX Support
Arize News
Arize Observe
Discussions
Introductions
Phoenix Support

Powered by Tightknit

Phoenix Support

Phoenix Support

Troubleshooting Azure OpenAI Model Request Timeout Issues

·Dec 11, 2024 10:57 PM

I'm trying out the prompt playground in the new hosted phoenix with azure openai models. After I specified the model config (endpoint, deployment name, api version) as well as the azure openai key, and try to run a prompt, i'm getting a request timeout after waiting a little bit.

8 comments

· Sorted by Oldest

RunLLM
APP
·
It seems likely that the issue you're experiencing with request timeouts when using Azure OpenAI models in Phoenix could be related to the configuration of the Azure OpenAI parameters. According to the Phoenix documentation, it's important to ensure that the azure_endpoint, azure_deployment, and api_version are correctly set. Additionally, the request_timeout parameter can be adjusted if the default timeout of 600 seconds is not sufficient.My understanding is that you should verify that all Azure-specific parameters are correctly configured and that the API key is valid. If the issue persists, consider increasing the request_timeout to allow more time for the request to complete. Would you like more detailed guidance on how to adjust these settings or troubleshoot further?Sources: - Phoenix documentation on Azure OpenAI
💡 Hint: Mention RunLLM in the thread for followups.
Mikyo
·
Hey Hain-Lee H. - it could be our security policy on the hosted phoenix. Can you let Kiko C. know the baseURL of azure?
Hain-Lee H.
·
Kiko C. it's https://openai-gpt4-internal-test.openai.azure.com/. We also have another one: https://openai-internal-test.openai.azure.com/
2
Kiko C.
·
Thanks Hain-Lee H., we’ll take a look and get back to you
🙌1
Hain-Lee H.
·
works now!
Hain-Lee H.
·
thanks!
Kiko C.
·
You got it, thanks for bringing it up!
Kiko C.
·
Enjoy!
🙌1