Cannot access to AML batch endpoints from Fabric pipelines

[francollado99]

I published a batch endpoint in Azure Machine Learning with the following URL 

https://name.eastus.inference.ml.azure.com/jobs

And when I try to access to it in a Web activity in Fabric I'm constantly getting the following 

Error calling the endpoint 'https://name.eastus.inference.ml.azure.com'. Response status code: 'NA - Unknown'. More details: Exception message: 'NA - Unknown [ClientSideException] An error occurred while sending the request.'.
Request didn't reach the server from the client. This could happen because of an underlying issue such as network connectivity, a DNS failure, a server certificate validation or a timeout. 

 Can I get some help? I can confirm that the AML workspace is open to all sources (no managed private endpoint needed). 

[V-yubandi-msft]

Hi @francollado99 ,

We haven’t received a response from your end yet. Please let us know whether the issue has been resolved or if you’re still facing any difficulties. Feel free to reach out if you need further assistance.

Thank you.

[V-yubandi-msft]

Hi @francollado99 ,

Please let us know if the issue is resolved or if you’re still facing any difficulties. Feel free to reach out if you need further assistance.

Thank You.

[francollado99]

No, It was not resolved. I prefered going back to pipelines v1 and old azureml libraries. Thanks.

[V-yubandi-msft]

Thanks for the update. No worries glad you were able to move forward using Pipelines v1.

If you ever want to revisit the Fabric Web activity approach, feel free to share the config details and we can take another look.

Appreciate you getting back to us.

[V-yubandi-msft]

Hi @francollado99 ,

Thank you for engaging with the Microsoft Fabric Community. Based on the error message, it looks like the Web activity request is not reaching the Azure ML endpoint. The issue seems to be happening on the Fabric side before the call is actually sent, which is why the status shows as NA – Unknown along with a client side exception.

Since the AML workspace is already open to public access, this usually points to a configuration or connectivity issue in the Web activity. Could you please check the following

1. Make sure the complete endpoint URL is being used, including the /jobs path and API version (if required).

2. Verify that the Authorization header is added correctly.

3. Try calling the same endpoint using Postman or a notebook to confirm whether it works outside Fabric.

4. If possible, test calling a simple public API from the same Web activity to validate outbound connectivity.

I hope this clarifies the situation. If I’ve misunderstood any part of your situation, please let us know.