clouder azure

Azure cloud operations — configure credentials, manage virtual machines, list resources and regions.

Setup

Before using Azure commands, you need to authenticate. Clouder supports two methods:

# Log in to Azure
az login

# Configure Clouder — auto-detects your az login session
clouder azure configure

The interactive flow:

Clouder asks if you want to use az login credentials
Fetches available subscriptions from your Azure account
Displays a table and lets you select one
Saves the subscription ID and tenant ID to ~/.clouder/clouds/azure/azure.yaml

Method 2: Service Principal (recommended for production/CI)

# Create a service principal
az ad sp create-for-rbac \
    --name "clouder-cli" \
    --role Contributor \
    --scopes "/subscriptions/<SUBSCRIPTION_ID>"

# Configure Clouder with the service principal
clouder azure configure \
    --subscription-id <SUBSCRIPTION_ID> \
    --tenant-id <TENANT_ID> \
    --client-id <CLIENT_ID> \
    --client-secret <CLIENT_SECRET> \
    --no-interactive

See the Azure Setup guide for detailed instructions on creating service principals.

Verify Configuration

# Show current Azure configuration
clouder azure

# List subscriptions
clouder azure subscriptions

Commands

`clouder azure configure`

Configure Azure credentials for Clouder.

# Interactive (default) — prompts for everything
clouder azure configure

# Non-interactive with flags
clouder azure configure \
    --subscription-id <ID> \
    --tenant-id <ID> \
    --client-id <ID> \
    --client-secret <SECRET> \
    --no-interactive

Option	Short	Description
`--subscription-id`	`-s`	Azure subscription ID
`--tenant-id`	`-t`	Azure tenant ID
`--client-id`	`-c`	Azure app (client) ID
`--client-secret`		Azure client secret
`--interactive / --no-interactive`		Enable/disable interactive prompts (default: interactive)

Configuration is stored in ~/.clouder/clouds/azure/azure.yaml with chmod 600.

`clouder azure subscriptions`

List all Azure subscriptions accessible by your credential.

clouder azure subscriptions

Output columns: Subscription ID, Name, State, Tenant ID.

`clouder azure regions`

List all available Azure regions/locations.

clouder azure regions

Output columns: Name (e.g., eastus), Display Name, Regional Name, Latitude, Longitude.

Example output:

Name	Display Name	Regional Name	Latitude	Longitude
eastus	East US	(US) East US	37.3719	-79.8164
westeurope	West Europe	(Europe) West Europe	52.3667	4.9000
southeastasia	Southeast Asia	(Asia Pacific) Southeast Asia	1.2830	103.8330

`clouder azure resource-groups`

List resource groups in the current subscription.

clouder azure resource-groups

Output columns: Name, Location, State, Tags.

`clouder azure resources`

List resources, optionally filtered by region or resource group.

# Interactive — prompts for region
clouder azure resources

# Filter by region
clouder azure resources --region eastus

# Filter by resource group
clouder azure resources --resource-group my-rg

Option	Short	Description
`--region`	`-r`	Filter by region (interactive if omitted)
`--resource-group`	`-g`	Filter by resource group

In interactive mode, Clouder shows a list of popular regions (eastus, westeurope, etc.) and lets you select one.

`clouder azure vm-sizes`

List available VM sizes in a region.

# Interactive — prompts for region
clouder azure vm-sizes

# Specific region
clouder azure vm-sizes --region westeurope

Option	Short	Description
`--region`	`-r`	Region to list VM sizes for (interactive if omitted)

Output columns: Name, Family, vCPUs, Memory (GB), Max Data Disks, Temp Disk (GB), OS Disk (GB), Arch, GPU, GPU Count, GPU Type, GPU Mem (GB).

Notes:

GPU is Yes when Azure SKU capabilities advertise GPU/accelerator support.
GPU Type and GPU Mem (GB) are shown when provided by Azure capability metadata; otherwise they appear as N/A.

Example output:

Name	Family	vCPUs	Memory (GB)	Max Data Disks	Temp Disk (GB)	OS Disk (GB)	Arch	GPU	GPU Count	GPU Type	GPU Mem (GB)
Standard_B2s	standardBSFamily	2	4.0	4	8.0	32	x64	No	0	N/A	N/A
Standard_D4s_v5	standardDSv5Family	4	16.0	8	75.0	32	x64	No	0	N/A	N/A
Standard_NC6s_v3	standardNCSv3Family	6	112.0	12	736.0	32	x64	Yes	1	NVIDIA Tesla V100	16.0

`clouder azure vm-ls`

List Azure virtual machines.

# All VMs in the subscription
clouder azure vm-ls

# Filter by resource group
clouder azure vm-ls --resource-group my-rg

Option	Short	Description
`--resource-group`	`-g`	Resource group to list VMs from

Output columns: Name, Location, VM Size, State, OS, Resource Group.

`clouder azure vm-create`

Create an Azure virtual machine. This is Clouder's primary command for provisioning cloud infrastructure.

Interactive mode (prompts for all values):

clouder azure vm-create

The interactive flow:

VM name — enter the name
Resource group — defaults to <name>-rg
Region — shows popular regions (eastus, westeurope, francecentral, etc.)
VM size — shows common sizes from free-tier to GPU:
- Standard_B1s — 1 vCPU, 1 GB RAM (free tier eligible)
- Standard_B2s — 2 vCPUs, 4 GB RAM
- Standard_B4ms — 4 vCPUs, 16 GB RAM
- Standard_D4s_v5 — 4 vCPUs, 16 GB RAM (compute optimized)
- Standard_D8s_v5 — 8 vCPUs, 32 GB RAM
- Standard_NC6s_v3 — 6 vCPUs, 112 GB RAM, 1 GPU V100
SSH key — auto-detects ~/.ssh/id_rsa.pub and asks to use it
Confirmation — shows a summary and asks to proceed

Non-interactive mode (all options specified):

clouder azure vm-create \
    --name my-vm \
    --resource-group my-rg \
    --region eastus \
    --vm-size Standard_B2s \
    --ssh-key ~/.ssh/id_rsa.pub \
    --image Ubuntu2204 \
    --tags "env=dev,team=data"

Option	Short	Default	Description
`--name`	`-n`		VM name
`--resource-group`	`-g`		Resource group (created if needed)
`--region`	`-r`		Azure region
`--vm-size`			VM size (e.g., `Standard_B2s`)
`--admin-user`		`azureuser`	Admin username
`--ssh-key`			Path to SSH public key file
`--image`		`Ubuntu2204`	OS image (see table below)
`--tags`			Comma-separated tags (`key=value,...`)

Available images:

Image	Publisher	OS
`Ubuntu2204`	Canonical	Ubuntu 22.04 LTS (default)
`Ubuntu2404`	Canonical	Ubuntu 24.04 LTS
`Debian12`	Debian	Debian 12

What gets created:

When you create a VM, Clouder automatically provisions:

A resource group (if it doesn't exist)
A virtual network and subnet (10.0.0.0/16)
A static public IP address
A network interface
The VM itself with SSH key authentication

After creation, Clouder shows the public IP and an SSH command:

┌─────────────── VM Created ───────────────┐
│  VM created successfully!                │
│                                          │
│  Name:             my-vm                 │
│  Location:         eastus                │
│  VM Size:          Standard_B2s          │
│  State:            Succeeded             │
│  Public IP:        20.xxx.xxx.xxx        │
│  Resource Group:   my-rg                 │
│                                          │
│  SSH:  ssh azureuser@20.xxx.xxx.xxx      │
└──────────────────────────────────────────┘

`clouder azure vm-delete`

Delete an Azure virtual machine.

# With confirmation prompt
clouder azure vm-delete my-vm --resource-group my-rg

# Skip confirmation
clouder azure vm-delete my-vm --resource-group my-rg --force

Argument / Option	Short	Description
`NAME` (required)		VM name to delete
`--resource-group` (required)	`-g`	Resource group containing the VM
`--force`	`-f`	Skip confirmation prompt

`clouder azure helm-values`

Generate a JSON values file containing all Azure credentials required by the datalayer-operator Helm chart:

operator.cloudCredentials.azure.tenantId
operator.cloudCredentials.azure.clientId
operator.cloudCredentials.azure.clientSecret
operator.cloudCredentials.azure.subscriptionId
operator.cloudCredentials.azure.resourceGroup

# Generate from configured Azure + kubeadm cluster metadata
clouder azure helm-values --cluster my-cluster -o datalayer-operator-azure.json

# Print JSON to stdout
clouder azure helm-values --cluster my-cluster --output -

# Disable service principal auto-creation (fail if client credentials are missing)
clouder azure helm-values --cluster my-cluster --no-create-sp

# Strict reuse mode (never creates an SP; fails if credentials are missing)
clouder azure helm-values --cluster my-cluster --reuse-sp-only

# Fully explicit mode
clouder azure helm-values \
    --tenant-id <TENANT_ID> \
    --client-id <CLIENT_ID> \
    --client-secret <CLIENT_SECRET> \
    --subscription-id <SUBSCRIPTION_ID> \
    --resource-group <RESOURCE_GROUP> \
    -o datalayer-operator-azure.json

Default output behavior:

If --cluster is set (or inferred from current kube context), output defaults to ~/.clouder/kubeadm/<cluster>/datalayer-operator-azure.json.
If no cluster can be resolved, output defaults to ~/.clouder/kubeadm/datalayer-operator-azure.json.

Option	Short	Description
`--cluster`		Kubeadm cluster name (reads `~/.clouder/kubeadm/<name>/kubeadm.json` for subscription/resource group)
`--output`	`-o`	Output JSON file path (`-` prints to stdout)
`--tenant-id`		Override tenant ID
`--client-id`		Override client ID
`--client-secret`		Override client secret
`--subscription-id`		Override subscription ID
`--resource-group`	`-g`	Override resource group
`--create-sp / --no-create-sp`		Auto-create scoped Azure service principal if client credentials are missing (default: `--create-sp`)
`--reuse-sp-only`		Strict mode: reuse existing credentials only and never create a service principal

Output file permissions are set to 600 because it contains secrets.

Use the generated file with Helm:

helm upgrade --install datalayer-operator \
    ./etc/helm/charts/datalayer-operator \
    --namespace datalayer-runtimes \
    --values datalayer-operator-azure.json

Example: Full Azure VM Workflow

# 1. Configure Azure (one-time)
clouder azure configure

# 2. Explore available regions and VM sizes
clouder azure regions
clouder azure vm-sizes --region eastus

# 3. Create a VM
clouder azure vm-create \
    --name jupyter-node-1 \
    --resource-group clouder-cluster-rg \
    --region eastus \
    --vm-size Standard_D4s_v5 \
    --ssh-key ~/.ssh/id_rsa.pub \
    --image Ubuntu2404 \
    --tags "role=jupyter,cluster=my-cluster"

# 4. SSH into the VM
ssh azureuser@<public-ip>

# 5. List your VMs
clouder azure vm-ls

# 6. Clean up
clouder azure vm-delete jupyter-node-1 \
    --resource-group clouder-cluster-rg

Setup​

Method 1: Azure CLI Login (recommended for development)​

Method 2: Service Principal (recommended for production/CI)​

Verify Configuration​

Commands​

clouder azure configure​

clouder azure subscriptions​

clouder azure regions​

clouder azure resource-groups​

clouder azure resources​

clouder azure vm-sizes​

clouder azure vm-ls​

clouder azure vm-create​

clouder azure vm-delete​

clouder azure helm-values​

Example: Full Azure VM Workflow​