Knowledge Graph — Coursera Notes › Academic disciplines › Information Technology / Computer Science › Artificial Intelligence › Machine Learning & Data › Azure Machine Learning Service

Azure ML authentication methods

feature · part of Azure Machine Learning Service

Azure Machine Learning supports multiple authentication methods to secure access to workspaces and resources. The five primary methods are:

• Interactive login authentication — default for development, uses Workspace.from_config().
• Azure CLI authentication — for CLI users, uses AzureCliAuthentication.
• Managed service identity (MSI) authentication — for Azure-hosted services, uses ManagedIdentityCredential.
• Service principal authentication — for automated workflows/CI/CD, uses ServicePrincipalAuthentication.
• Token authentication — for external integrations, uses InteractiveLoginAuthentication with a token.

Each method has specific use cases and setup steps.

For troubleshooting, specify Subscription ID, Resource group, and Workspace name from Azure portal.

• Azure CLI authentication — preferred for users comfortable with CLI tools. Steps: 1. Install Azure CLI. 2. Log in via az login. 3. In script: from azureml.core.authentication import AzureCliAuthentication then cli_auth = AzureCliAuthentication() and ws = Workspace.from_config(auth=cli_auth).

az login
from azureml.core.authentication import AzureCliAuthentication
cli_auth = AzureCliAuthentication()
ws = Workspace.from_config(auth=cli_auth)

• Managed identity authentication — eliminates passwords/secrets, highly secure for Azure-hosted services. Steps: 1. Ensure environment supports MSI (Azure VMs/services). 2. Use ManagedIdentityCredential from azure.identity. 3. Create workspace: ws = Workspace(subscription_id=..., resource_group=..., workspace_name=..., credential=credential).

from azure.identity import ManagedIdentityCredential
from azureml.core import Workspace
credential = ManagedIdentityCredential()
ws = Workspace(subscription_id="your_subscription_id", resource_group="your_resource_group", workspace_name="your_workspace_name", credential=credential)

• Service principal authentication — ideal for automated workflows/CI/CD with strict access control. Steps: 1. Register an app in Azure AD, record Application ID, Tenant ID, Client Secret. 2. Grant Service Principal access to Azure ML workspace. 3. In script: from azureml.core.authentication import ServicePrincipalAuthentication then svc_pr = ServicePrincipalAuthentication(tenant_id=..., service_principal_id=..., service_principal_password=...) and ws = Workspace.from_config(auth=svc_pr).

from azureml.core.authentication import ServicePrincipalAuthentication
svc_pr = ServicePrincipalAuthentication(tenant_id="your_tenant_id", service_principal_id="your_application_id", service_principal_password="your_client_secret")
ws = Workspace.from_config(auth=svc_pr)

• Token authentication — offers granular control for external integrations. Steps: 1. Generate a token externally using Azure-compatible tools. 2. In script: from azureml.core.authentication import InteractiveLoginAuthentication then token_auth = InteractiveLoginAuthentication(token="your_token_here") and ws = Workspace.from_config(auth=token_auth).

from azureml.core.authentication import InteractiveLoginAuthentication
token_auth = InteractiveLoginAuthentication(token="your_token_here")
ws = Workspace.from_config(auth=token_auth)

• Key Vault authentication — securely manage secrets (API keys, passwords) in remote runs. Steps: 1. Link Azure ML workspace to a Key Vault. 2. Add secrets as key-value pairs. 3. Access in script: ws = Workspace.from_config() then secret = ws.get_default_keyvault().get_secret(name="your_secret_name").

from azureml.core import Workspace
ws = Workspace.from_config()
secret = ws.get_default_keyvault().get_secret(name="your_secret_name")

Also known as: Azure ML authentication methods, ML authentication methods

This is the text view of an interactive 3D knowledge graph — open this page with JavaScript enabled to explore it visually.