Accelerator guide ADO with Terraform misses creation of dev AML environment

[corticalstack]

Describe the bug or the issue that you are facing

Hi,

Following the accelerator guide for ADO and Terraform, I have successfully competed the steps to provision the prod env resources. This is via a pipeline that executes /infrastructure/devops-pipelines/tf-ado-deploy-infra.yml from the main branch.

There seems to be an omission in the guide for provisioning the dev environment?

Steps/Code to Reproduce

As described.

Expected Output

As described.

Versions

As described.

Which platform are you using for deploying your infrastrucutre?

Azure DevOps (ADO)

If you mentioned Others, please mention which platformm are you using?

No response

What are you using for deploying your infrastrucutre?

Terraform

Are you using Azure ML CLI v2 or Azure ML Python SDK v2

Azure ML Python SDK v2

Describe the example that you are trying to run?

As described.

[setuc]

@corticalstack This is a manual step that's required. We can add steps on how to create the environment in Azure DevOps under step 3.10 in the documentation. Or if folks have the CLI installed they can use az pipelines environment create to create the dev environment.

[setuc]

@sdonohoo Marking this step for us to add step to create the environment in step 3.10

[corticalstack]

@sdonohoo Marking this step for us to add step to create the environment in step 3.10

It would certainly make sense to have the dev environment creation documented, especially given there is code for it in tf-ado-deploy-infra.yml and the need to update config-infra-dev.yml.

Thank-you in advance, and for all your hard work on this to date.

[corticalstack]

Having walked through the accelerator outer-inner loop steps, I'd say what currently exists very much focuses on a single environment setup (creating environment, code copied/missed when setting up a 2-env landscape, which branch to use when running pipeline) which kinda misses the point of MLops and automated packaging of models from dev to prod. A thorough and accurate 2 end step-by step would be super useful. With Python SDK v2 would be a bonus :)

[setuc]

@corticalstack Thank you for taking time to share your feedback and log the issues. These types of feedback are invaluable to us. Would you be open for a chat on what you think the ideal scenario would be? We are planning our next build and would love to hear more of your ideas and suggestions on how we can make this MLOps project even better. I am reachable via the social media links on my profile, or you can suggest a way to get in touch.