HTTP error while calling OpenAI models using Microsoft Foundry

Question

HTTP error while calling OpenAI models using Microsoft Foundry

Agarwal, Anjali 20

Code.py %pip install azure-ai-projects==2.0.0b2 openai==1.109.1 python-dotenv azure-identity

import os from dotenv import load_dotenv from azure.identity import DefaultAzureCredential from azure.ai.projects import AIProjectClient load_dotenv() foundry_project_endpoint = os.getenv("FOUNDRY_PROJECT_ENDPOINT") model_deployment_name = os.getenv("MODEL_DEPLOYMENT_NAME")

project_client = AIProjectClient( endpoint= foundry_project_endpoint, credential=DefaultAzureCredential() )

openai_client = project_client.get_openai_client()

response = openai_client.responses.create( model=model_deployment_name, instructions="You are a helpful AI assistant.", input = "Can you tell me about Microsoft Foundry?" ) print(f"Response output: {response.output_text}")

And .env has:

FOUNDRY_PROJECT_ENDPOINT="https://demoprojectv1.services.ai.azure.com/api/projects/udemy-demo-proj" MODEL_DEPLOYMENT_NAME="gpt-4o"

Error:

ConnectError Traceback (most recent call last) File ~\AppData\Roaming\Python\Python313\site-packages\httpx_transports\default.py:101, in map_httpcore_exceptions() 100 try: --> 101 yield 102 except Exception as exc:

File ~\AppData\Roaming\Python\Python313\site-packages\httpx_transports\default.py:250, in HTTPTransport.handle_request(self, request) 249 with map_httpcore_exceptions(): --> 250 resp = self._pool.handle_request(req) 252 assert isinstance(resp.stream, typing.Iterable)

File ~\AppData\Roaming\Python\Python313\site-packages\httpcore_sync\connection_pool.py:256, in ConnectionPool.handle_request(self, request) 255 self._close_connections(closing) --> 256 raise exc from None 258 # Return the response. Note that in this case we still have to manage 259 # the point at which the response is closed.

File ~\AppData\Roaming\Python\Python313\site-packages\httpcore_sync\connection_pool.py:236, in ConnectionPool.handle_request(self, request) 234 try: 235 # Send the request on the assigned connection. --> 236 response = connection.handle_request( 237 pool_request.request 238 ) 239 except ConnectionNotAvailable: ... (...) 1022 response.headers, 1023 ) 1024 log.debug("request_id: %s", response.headers.get("x-request-id"))

APIConnectionError: Connection error. Output is truncated. View as a scrollable element or open in a text editor. Adjust cell output settings...

Anshika Varshney 10,315 Reputation points Microsoft External Staff Moderator

2026-04-24T02:08:19.7866667+00:00
Hello Anjali,Thanks for sharing the details in the private chat.

Your endpoint and DNS look correct, so this is not a network issue. The problem is mainly related to the SDK and method usage.

The error AIProjectClient has no attribute get_openai_client usually means the installed SDK version does not support this method or there is a version mismatch. The method is only available in specific versions of the Azure AI Projects SDK. [learn.microsoft.com]

Please check these points:

Verify the installed SDK version and upgrade to the latest version

Restart your environment after upgrading

Confirm the method exists by checking available functions in the client

Also double check:

Use deployment name, not model name

Ensure the model is deployed and active in your Foundry project

In short, your setup looks fine, but this issue is caused by SDK version mismatch or incorrect method support, not networking.

I Hope this helps. Do let me know if you have any further queries.
Thankyou!
SRILAKSHMI C 18,205 Reputation points Microsoft External Staff Moderator

2026-04-24T17:25:02.1266667+00:00

Hi Agarwal, Anjali

Thank you for the update.

I’m glad to hear that you were able to resolve the issue. Installing pip_system_certs makes sense in this scenario, as it enables Python to use the system’s trusted certificate store for TLS/SSL validation.

I appreciate you sharing the resolution it may also help in similar situations going forward.

Since I’ve converted my earlier comment into an answer, could you please take a moment to mark it as Accepted? This helps others in the community with the same question find the solution more easily.

Thank you

Answer accepted by question author

2 additional answers

Your answer

Anshika Varshney 10,315 Reputation points Microsoft External Staff Moderator

2026-04-24T02:08:19.7866667+00:00

Hello Anjali,Thanks for sharing the details in the private chat.

Your endpoint and DNS look correct, so this is not a network issue. The problem is mainly related to the SDK and method usage.

The error AIProjectClient has no attribute get_openai_client usually means the installed SDK version does not support this method or there is a version mismatch. The method is only available in specific versions of the Azure AI Projects SDK. [learn.microsoft.com]

Please check these points:

Verify the installed SDK version and upgrade to the latest version

Restart your environment after upgrading

Confirm the method exists by checking available functions in the client

Also double check:

Use deployment name, not model name

Ensure the model is deployed and active in your Foundry project

In short, your setup looks fine, but this issue is caused by SDK version mismatch or incorrect method support, not networking.

I Hope this helps. Do let me know if you have any further queries.
Thankyou!
SRILAKSHMI C 18,205 Reputation points Microsoft External Staff Moderator

2026-04-24T17:25:02.1266667+00:00

Hi Agarwal, Anjali

Thank you for the update.

I’m glad to hear that you were able to resolve the issue. Installing pip_system_certs makes sense in this scenario, as it enables Python to use the system’s trusted certificate store for TLS/SSL validation.

I appreciate you sharing the resolution it may also help in similar situations going forward.

Since I’ve converted my earlier comment into an answer, could you please take a moment to mark it as Accepted? This helps others in the community with the same question find the solution more easily.

Thank you

Answer 1

Hello @Agarwal, Anjali

Thank you for sharing the code snippet and detailed error message.

The APIConnectionError / ConnectError you’re encountering indicates that the request is failing at the network/connection layer, meaning it is not reaching the Azure AI Foundry (OpenAI) service. This is typically related to endpoint configuration, authentication, or network restrictions, rather than an issue with the model or SDK usage.

Your code structure and SDK usage look correct. Based on the details provided, the issue is most likely caused by one or more of the following:

Endpoint not reachable or incorrectly formatted
Authentication not being resolved via DefaultAzureCredential
Network/firewall/proxy restrictions
Environment variable formatting issues

Recommended Troubleshooting Steps

1. Validate the Foundry Project Endpoint

Ensure the endpoint format is correct:

https://<your-resource>.services.ai.azure.com/api/projects/<your-project-id>

In your case:

https://demoprojectv1.services.ai.azure.com/api/projects/udemy-demo-proj

Also, update your .env file to remove quotes:

FOUNDRY_PROJECT_ENDPOINT=https://demoprojectv1.services.ai.azure.com/api/projects/udemy-demo-proj

Then verify it is reachable:

Open in browser:

https://demoprojectv1.services.ai.azure.com

Or test via curl:

curl https://demoprojectv1.services.ai.azure.com

If this fails, it indicates a DNS or network connectivity issue.

2. Verify Environment Variable Loading

Confirm values are being read correctly in your script:

print(foundry_project_endpoint)

Ensure there are no extra spaces or hidden characters.

3. Confirm Authentication

DefaultAzureCredential() requires a valid identity. Please ensure one of the following:

Login via Azure CLI:

az login

OR set service principal credentials:

AZURE_CLIENT_ID

If authentication is not properly established, the SDK may fail during connection initialization.

4. Verify Azure Permissions

Ensure the identity being used has:

Reader role on the Foundry project
Cognitive Services OpenAI User role on the underlying Azure OpenAI resource

5. Check Network / Firewall / Proxy

If you are on a corporate network:

Ensure outbound HTTPS (port 443) is allowed

Allow access to:

*.services.ai.azure.com

If using a proxy, configure:

set HTTPS_PROXY=http://<proxy>:<port>

You can also test connectivity:

nslookup demoprojectv1.services.ai.azure.com

6. SDK and Python Version Compatibility

You are using:

azure-ai-projects==2.0.0b2 (preview SDK)
Python 3.13

Some Azure SDKs may have limited validation on newer Python versions.

Recommendation:

Try with Python 3.10 or 3.11
Upgrade SDK if a newer version is available

7. Isolate the Issue

Test basic client initialization:

from azure.identity import DefaultAzureCredential

If this fails, it confirms the issue is before the OpenAI API call (connection/auth level).

Please refer this

Troubleshoot endpoint and model discovery (Cannot find model): https://learn.microsoft.com/azure/ai-foundry/reference/region-support

HTTP 404 “Operation Not Found” (ensure correct endpoint & deployment name): https://learn.microsoft.com/azure/ai-foundry/openai/latest

I hope this will help you. Please feel free to let me know if you have any other queries.

Thank you!

Answer 2

Agarwal, Anjali 20

Hi Anshika,

I was working with Capgemini's internal support team and they helped me with a solution that worked.

All I had to do was - pip install pip_system_certs

Somehow Python was unable to verify TLS/SSL connections to servers whose certificates are trusted by my system.

So, you may close this case now. Thank you so much for being patient with me and helping me all this while!

0 comments

Answer 3

Thanks for your reply, Anshika.

I have tried multiple combinations in the last few days and none worked for me. Can you kindly rewrite the below code with the correct SDK and then may be I can try running it?

Code.py %pip install azure-ai-projects==2.0.0b2 openai==1.109.1 python-dotenv azure-identity

import os from dotenv import load_dotenv from azure.identity import DefaultAzureCredential from azure.ai.projects import AIProjectClient load_dotenv() foundry_project_endpoint = os.getenv("FOUNDRY_PROJECT_ENDPOINT") model_deployment_name = os.getenv("MODEL_DEPLOYMENT_NAME")

project_client = AIProjectClient( endpoint= foundry_project_endpoint, credential=DefaultAzureCredential() )

openai_client = project_client.get_openai_client()

response = openai_client.responses.create( model=model_deployment_name, instructions="You are a helpful AI assistant.", input = "Can you tell me about Microsoft Foundry?" ) print(f"Response output: {response.output_text}")

And .env has:

FOUNDRY_PROJECT_ENDPOINT="https://demoprojectv1.services.ai.azure.com/api/projects/udemy-demo-proj" MODEL_DEPLOYMENT_NAME="gpt-4o"

Share via

HTTP error while calling OpenAI models using Microsoft Foundry

Error:

2 additional answers

Your answer