How can we improve Microsoft Azure Machine Learning?

Add FQDN as a parameter in Azure ML Webservice.deploy_from_image/model

An Azure ML web service deployed to ACI gives you an IP address to request response. However, this IP address can change on its own because of some Azure Internal Events on ACI.

The Azure Container Instance (ACI) allows you to define a FQDN which binds to the underlying IP address so that you don't have to deal with the IP Address. However, it is proving impossible to update the Azure ML deployed ACI because of various issues. If FQDN can be added as a parameter in Webservice.deploy_from_image which binds to the ACI container's FQDN, this would solve all the problem.

88 votes
Vote
Sign in
(thinking…)
Sign in with: oidc
Signed in as (Sign out)
You have left! (?) (thinking…)
Karan Joshi shared this idea  ·   ·  Flag idea as inappropriate…  ·  Admin →

1 comment

Sign in
(thinking…)
Sign in with: oidc
Signed in as (Sign out)
Submitting...
  • Salil Bapat commented  ·   ·  Flag as inappropriate

    We have fixed this by having an FQDN on the service deployed to ACI. Your new deployments for the same service will get a endpoint with FQDN rather than IP address. The SDK change where you can choose the name will get rolled out in next few days.

Feedback and Knowledge Base