Clarification requested on creating endpoints with serverless compute

[ma261065]

This document mentions how to set up ACR to enable a private endpoint.

But there should be a slight change here: AML by default now uses serverless compute to build the docker image since one or two months ago.

So if the ACR is private, serverless compute can't access customer's ACR and the build image will fail.

Even though this document mentions that they can create a compute cluster for building the image, some customers complain they will be charged for storage even the compute cluster is idle.

I suppose that's why we changed to use serverless compute for image build, so that customers pay only when it is used.

Therefore, I think we should add a note in 5. that AML by default uses serverless compute for the image build and suggest that a customer can also configure serverless compute in their owned vnet to avoid this issue.

---

Document Details

⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.

• ID: 9589d2d0-2322-3106-c96b-982bd8296594
• Version Independent ID: 887f36ae-7270-e712-eb81-fcfe53d8a164
• Content: Secure an Azure Machine Learning workspace with virtual networks - Azure Machine Learning
• Content Source: articles/machine-learning/how-to-secure-workspace-vnet.md
• Service: machine-learning
• Sub-service: enterprise-readiness
• GitHub Login: @jhirono
• Microsoft Alias: jhirono

[TPavanBalaji]

@ma261065 Thanks for your feedback! We will investigate and update as appropriate.

[AjayBathini-MSFT]

@ma261065 Thanks for bringing this to our attention. I'm going to assign this to the document author so they can take a look at it accordingly.

[jhirono]

@Blackmist could you please update the doc accordingly? I agree with what OP said.

[Blackmist]

@ma261065 thanks for the feedback! We published an update to the article yesterday to call out the serverless compute scenario.

please-close