Azure ML Managed Endpoint Blocks Port 80 - Model Fails to Connect Back

[ishu-harsh]

Describe the bug

I deployed a Dockerized model on an Azure ML GPU instance using AML Managed Endpoints. However, the application inside the container is trying to make requests on port 80, which AML blocks by default since it only allows HTTPS (port 443). This results in connection failures and API timeouts.

To Reproduce

Steps to reproduce the behavior:

Deploy Omniparse docker image inside an Azure ML GPU Compute Instance using AML Managed Endpoints.

The application tries to connect to:

http://jay1-8000.italynorth.instances.azureml.ms/parse_image/process_image

Since AML blocks port 80, the request times out and fails.

The following error is logged:

gradio.exceptions.Error: "Failed to parse: HTTPConnectionPool(host='jay1-8000.italynorth.instances.azureml.ms', port=80): 
Max retries exceeded with url: /parse_image/process_image 
(Caused by ConnectTimeoutError(<urllib3.connection.HTTPConnection object at 0x7fd8da27a320>, 
'Connection to jay1-8000.italynorth.instances.azureml.ms timed out. (connect timeout=None)'))"

Expected behavior

The application should be able to make internal API requests within AML Compute, even if the external endpoint requires HTTPS.
Either AML should allow internal HTTP routing, or there should be an official workaround for internal API calls.

Screenshots
If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information):

• OS: [e.g. iOS]
• Browser [e.g. chrome, safari]
• Version [e.g. 22]

Smartphone (please complete the following information):

• Device: [e.g. iPhone6]
• OS: [e.g. iOS8.1]
• Browser [e.g. stock browser, safari]
• Version [e.g. 22]

Additional context

Questions:

Is there a way to allow internal HTTP requests within AML Compute?
Can internal API calls be automatically routed via HTTPS?
Would switching to an Azure GPU VM instance be the only solution?

Would appreciate any guidance. Thanks in advance!