Azure Machine Learning (+Synapse) Example

This repo shows an example for rolling out a complete Azure Machine Learning and Azure Synapse Analytics (optional) enterprise environment via Terraform.

[TBC add diagram]

This includes rollout of the following resources:

• Azure Synapse Analytics Workspace with Private Endpoints
• Azure Synapse Analytics Private Link Hub with Private Endpoint
• Azure Storage Account with Private Endpoints for blob and dfs
• Azure Machine Learning Workspace with Private Endpoint
• Azure Key Vault with Private Endpoint
• Azure Container Registry with Private Endpoint
• Application Insights
• Virtual Network
• Jumphost (Windows) with Bastion for easy access to the VNet

Instructions

Make sure you have the Azure CLI and Terraform installed.

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

Notes

See notes below for additional info:

• A Synapse workspace is deployed by default, change enable_syn_ws value if you wanna disable it.
• A public IP is added to Azure Storage Account and Azure Synapse Analytics firewall rules to enable the deployment. That rule could be removed once the deployment is finished, only limiting jumphost access.
• Change enable_syn_sqlpool and enable_syn_sparkpool values if you wanna deploy any of those pools.
• Change enable_aml_computeinstance value if you wanna deploy a compute instance.
• The deployment was tested on wsl (ubuntu).