In the the section #using-a-workspace-over-a-private-endpoint on the page - where it says "communication to the workspace is only allowed from the virtual network" it doesn't specifically mention whether the API endpoints underpinning the Azure Machine Learning SDK for Python are also covered by this protection. Could you confirm and add that to this document please?
⚠Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.
To add to @rohancragg 's question.
I would be interested to know if this would have an impact on the consumption of AzureML pipelines.
Presently, it is my understanding that a PublishedPipeline is given a public facing endpoint.
For example, https://westeurope.api.azureml.ms/pipelines/v1.0/..../PipelineSubmit/
Would this pattern remain the same under when considering Azure Private LInk?
@rohancragg , @rohancragg
Thank you for reaching out. At this time we are reviewing the ask and will provide an update as appropriate
@aashishb Hi, could you please check this issue and update the document if necessary? Thanks a lot.
@jhirono ,can you check?
Hi @rohancragg , thanks for asking about this. I've spoken to the engineering team and they have confirmed that the APIs and SDK endpoints are covered by the private endpoint. I'll work with engineering to update the document with this information.
Hi @ernieglink , thanks for asking about the ml pipeline endpoint. From speaking to engineering, communication to the published pipeline should also be covered by the private endpoint. I'll work with engineering to update the documentation with information on this.
The doc update has been merged and will go live later today.