Azure CLI

Warnings

• gotcha The `azure-cli` PyPI package is the source for the command-line tool, but end-users are strongly advised to install it via official installers (e.g., `apt`, `yum`, `brew`, MSI) or the install script, rather than `pip install`. `pip install azure-cli` is primarily for developers, CI/CD, or highly isolated environments where pathing and dependencies must be managed carefully, and may not automatically set up the `az` executable on the system PATH.
  Fix: Refer to the official Azure CLI installation guide for your operating system (aka.ms/installazurecliwindows, aka.ms/installazureclinux, aka.ms/installazurecliosx).
• breaking Azure CLI continuously updates its Python version requirements. Starting with 2.49.0, Python >=3.10 is required. For 2.61.0+, Python >=3.11 is required. Newer versions have incrementally dropped support for older Python versions (e.g., Python 3.7 support dropped in 2.40.0, Python 3.8 dropped for versions requiring 3.10+). Ensure your Python environment meets the minimum requirement for your Azure CLI version.
  Fix: Upgrade your Python environment to meet the minimum requirement for the desired Azure CLI version. Check `az --version` for current Python runtime. If using `pip install`, ensure your virtual environment uses a compatible Python version.
• gotcha Azure CLI is a command-line tool, not a Python SDK. Do not attempt to `import azure.cli` to interact with Azure resources from your Python application. For programmatic interaction with Azure services, use the dedicated `azure-sdk-for-python` libraries (e.g., `azure-mgmt-storage`, `azure-identity`, `azure-storage-blob`). The `azure-cli` components are for extending the CLI itself or for internal operations.
  Fix: For Python development, use the `azure-sdk-for-python` client libraries relevant to the Azure service you wish to interact with. If you need to execute CLI commands from Python, use `subprocess.run(['az', ...])`.
• deprecated The `--json` output option has been deprecated in favor of `--output json` (or `-o json`). While `--json` might still work for some commands, `az` strongly recommends using the standard `--output` or `-o` parameter for consistency.
  Fix: Always use `--output json` or `-o json` when specifying JSON output format for Azure CLI commands.
• breaking The `az feedback` command was removed in Azure CLI 2.50.0. This command was used to provide feedback directly to the Azure CLI team.
  Fix: To provide feedback, use the GitHub issues page (https://github.com/Azure/azure-cli/issues) or other official feedback channels.

Install

• pip install azure-cli For development or isolated environments
• curl -sL https://aka.ms/InstallAzureCLIDeb | sudo bash Recommended for Linux (Debian/Ubuntu)
• brew update && brew install azure-cli Recommended for macOS

Imports

• azure.cli.core

  from azure.cli.core import get_default_cli

  The `azure-cli` package is not designed for direct import and use as a typical Python SDK. Its modules (like `azure.cli.core` or `azure.cli.command_modules`) are primarily for extending the CLI itself or for internal use. For interacting with Azure services from Python, use the specific `azure-sdk-for-python` client libraries (e.g., `azure-mgmt-compute`, `azure-storage-blob`).

Quickstart

This quickstart demonstrates how to execute Azure CLI commands programmatically from Python using the `subprocess` module. Users must have the Azure CLI installed and authenticated (e.g., via `az login`) prior to running these commands. This approach is suitable for scripting automation or integrating CLI operations within Python applications.

import subprocess
import os

# Ensure Azure CLI is installed and user is logged in (e.g., via `az login` in shell)
# You can also use service principal authentication for CI/CD environments.

def run_az_command(command):
    """Helper to run an az CLI command."""
    full_command = ['az'] + command.split()
    try:
        result = subprocess.run(full_command, capture_output=True, text=True, check=True, env=os.environ)
        print(f"Command: {' '.join(full_command)}")
        print(f"STDOUT:\n{result.stdout}")
        if result.stderr:
            print(f"STDERR:\n{result.stderr}")
        return result.stdout
    except subprocess.CalledProcessError as e:
        print(f"Error running command: {' '.join(full_command)}")
        print(f"Error STDOUT:\n{e.stdout}")
        print(f"Error STDERR:\n{e.stderr}")
        raise

# Example 1: List resource groups in JSON format
print("\n--- Listing resource groups ---")
resource_groups_json = run_az_command("group list --output json")
# Further processing with resource_groups_json if needed, e.g., json.loads()

# Example 2: Get current account details
print("\n--- Getting current account details ---")
account_details = run_az_command("account show --output tsv")

# Example 3: Create a resource group (requires appropriate permissions)
# Replace with unique name if running interactively
# RG_NAME = os.environ.get('AZURE_TEST_RG', 'my-python-test-rg')
# LOCATION = os.environ.get('AZURE_TEST_LOCATION', 'eastus')
# print(f"\n--- Creating resource group {RG_NAME} in {LOCATION} ---")
# try:
#     run_az_command(f"group create --name {RG_NAME} --location {LOCATION} --output none")
#     print(f"Resource group '{RG_NAME}' created successfully.")
# except Exception:
#     print(f"Could not create resource group '{RG_NAME}'. It might already exist or permissions are insufficient.")

# Example 4: Delete a resource group (USE WITH CAUTION!)
# print(f"\n--- Deleting resource group {RG_NAME} ---")
# run_az_command(f"group delete --name {RG_NAME} --no-wait --yes")
# print(f"Deletion command issued for resource group '{RG_NAME}'.")