Skip to content

An example repo for provisioning a complete Azure Machine Learning environment through Terraform

License

Notifications You must be signed in to change notification settings

lxyea/azure-machine-learning-terraform

 
 

Repository files navigation

Azure Machine Learning Enterprise Terraform Example

This repo shows an example for rolling out a complete Azure Machine Learning enterprise enviroment via Terraform.

Deployed resources

This includes rollout of the following resources:

  • Azure Machine Learning Workspace with Private Link
  • Azure Storage Account with VNET binding (using Service Endpoints) and Private Link for Blob and File
  • Azure Key Vault with VNET binding (using Service Endpoints) and Private Link
  • Azure Container Registry
  • Azure Application Insights
  • Virtual Network
  • Jumphost (Windows) with Bastion for easy access to the VNET
  • Compute Cluster (in VNET)
  • Compute Instance (in VNET)
  • (Azure Kubernetes Service - disabled by default and still under development)

Instructions

Make sure you have the Azure CLI and the Azure Machine Learning CLI extension installed (az extension add -n azure-cli-ml).

  1. Copy terraform.tfvars.example to terraform.tfvars
  2. Update terraform.tfvars with your desired values
  3. Run Terraform
    $ terraform init
    $ terraform plan
    $ terraform apply

If you want to deploy AKS, you need to have Azure Machine Learning CLI installed.

Known Limitations

  • Still need to update Default Datastore to use Managed Identity of the Studio UI data access

Important Notes

  • The user fileshare for the Compute Instances will be automatically provisioned upon first instance access

About

An example repo for provisioning a complete Azure Machine Learning environment through Terraform

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • HCL 100.0%