Azure Container Instances

Azure Container Instances (ACI) provides fast deployment of Linux containers without managing virtual machines. Deploy containers with public IP addresses, custom resource allocations, and private registry support.

List Containers

Retrieve all Azure Container Instances across subscriptions and resource groups.

Endpoint

GET /api/azure/containers

Authentication

Requires access_token in session.

Response

{
  "value": [
    {
      "name": "my-container",
      "resourceGroup": "my-rg",
      "location": "westeurope",
      "status": "Running",
      "image": "nginx:latest",
      "subscriptionId": "12345678-1234-1234-1234-123456789abc"
    }
  ]
}

Implementation

The list operation iterates through all subscriptions and resource groups:

from azure.mgmt.containerinstance import ContainerInstanceManagementClient
from azure.mgmt.resource import ResourceManagementClient
from azure.mgmt.subscription import SubscriptionClient

def list_containers():
    credential = FlaskCredential()
    sub_client = SubscriptionClient(credential)
    
    items = []
    for sub in sub_client.subscriptions.list():
        subscription_id = sub.subscription_id
        container_client = ContainerInstanceManagementClient(
            credential, subscription_id
        )
        resource_client = ResourceManagementClient(
            credential, subscription_id
        )
        
        for rg in resource_client.resource_groups.list():
            for cn in container_client.container_groups.list_by_resource_group(rg.name):
                details = container_client.container_groups.get(rg.name, cn.name)
                items.append({
                    "name": cn.name,
                    "resourceGroup": rg.name,
                    "location": cn.location,
                    "status": details.instance_view.state,
                    "image": cn.containers[0].image if cn.containers else "unknown",
                    "subscriptionId": subscription_id
                })
    
    return jsonify({"value": items})

Create Container

Deploy a new container instance with custom configurations.

Endpoint

POST /api/azure/containers

Request Body

{
  "subscriptionId": "12345678-1234-1234-1234-123456789abc",
  "rgName": "my-resource-group",
  "location": "westeurope",
  "cnName": "my-container",
  "image": "nginx:latest",
  "cpu": 1.0,
  "memory": 1.5,
  "port": 80,
  "registry": {
    "server": "myregistry.azurecr.io",
    "username": "myusername",
    "password": "mypassword"
  }
}

Parameters

subscriptionId

string

required

Azure subscription ID where the container will be created

rgName

string

required

Resource group name for the container

location

string

default:"westeurope"

Azure region for deployment (e.g., westeurope, eastus)

cnName

string

required

Container group name

image

string

required

Container image (e.g., nginx:latest, myregistry.azurecr.io/myapp:v1)

cpu

number

default:"1.0"

Number of CPU cores (0.1 to 4.0)

memory

number

default:"1.5"

Memory allocation in GB (0.1 to 16.0)

port

integer

default:"80"

Container port to expose

registry

object

Private registry credentials (optional)

server: Registry server URL
username: Registry username
password: Registry password

Response

{
  "message": "Kontener 'my-container' utworzony pomyślnie",
  "containerId": "/subscriptions/.../resourceGroups/my-rg/providers/Microsoft.ContainerInstance/containerGroups/my-container"
}

Implementation

from azure.mgmt.containerinstance.models import (
    ContainerGroup,
    Container,
    ResourceRequests,
    ResourceRequirements,
    OperatingSystemTypes,
    IpAddress,
    Port,
    ContainerPort,
    ContainerGroupNetworkProtocol,
    ImageRegistryCredential
)

def create_container():
    data = request.get_json()
    subscription_id = data.get("subscriptionId")
    rg_name = data.get("rgName")
    location = data.get("location", "westeurope")
    cn_name = data.get("cnName")
    image = data.get("image")
    cpu = data.get("cpu", 1.0)
    memory = data.get("memory", 1.5)
    port = data.get("port", 80)
    registry = data.get("registry")
    
    credential = FlaskCredential()
    client = ContainerInstanceManagementClient(credential, subscription_id)
    
    # Define resource requirements
    container_resource_requests = ResourceRequests(
        cpu=cpu,
        memory_in_gb=memory
    )
    container_resource_requirements = ResourceRequirements(
        requests=container_resource_requests
    )
    
    # Configure container
    container = Container(
        name=cn_name,
        image=image,
        resources=container_resource_requirements,
        ports=[ContainerPort(port=port)]
    )
    
    # Configure public IP address
    ip_address = IpAddress(
        ports=[Port(protocol=ContainerGroupNetworkProtocol.TCP, port=port)],
        type="Public"
    )
    
    # Add registry credentials if provided
    registry_credentials = []
    if registry:
        registry_credentials.append(ImageRegistryCredential(
            server=registry.get("server"),
            username=registry.get("username"),
            password=registry.get("password")
        ))
    
    # Create container group
    group = ContainerGroup(
        location=location,
        containers=[container],
        os_type=OperatingSystemTypes.LINUX,
        restart_policy="Always",
        ip_address=ip_address,
        image_registry_credentials=registry_credentials
    )
    
    result = client.container_groups.begin_create_or_update(
        resource_group_name=rg_name,
        container_group_name=cn_name,
        container_group=group
    ).result()
    
    return jsonify({
        "message": f"Kontener '{cn_name}' utworzony pomyślnie",
        "containerId": result.id
    }), 200

Container Configuration Details

Resource Allocation:

CPU cores are specified as floating-point numbers (e.g., 0.5, 1.0, 2.0)
Memory is specified in gigabytes
Azure charges based on allocated resources, not actual usage

Operating System:

Only Linux containers are supported in this implementation
Windows containers require OperatingSystemTypes.WINDOWS

Restart Policy:

Always: Restart container on failure (default)
OnFailure: Restart only if container exits with error
Never: Never restart the container

Network Configuration:

Public IP addresses are automatically assigned
Multiple ports can be exposed
Supports TCP and UDP protocols

Restart Container

Restart a running container instance.

Endpoint

POST /api/azure/containers/restart

Request Body

{
  "subscriptionId": "12345678-1234-1234-1234-123456789abc",
  "resourceGroup": "my-rg",
  "containerName": "my-container"
}

Parameters

subscriptionId

string

required

Azure subscription ID

resourceGroup

string

required

Resource group name

containerName

string

required

Container group name to restart

Response

{
  "message": "Kontener 'my-container' został zrestartowany"
}

Implementation

def restart_container():
    data = request.get_json()
    subscription_id = data.get("subscriptionId")
    resource_group = data.get("resourceGroup")
    container_name = data.get("containerName")
    
    credential = FlaskCredential()
    client = ContainerInstanceManagementClient(credential, subscription_id)
    
    client.container_groups.begin_restart(
        resource_group,
        container_name
    ).wait()
    
    return jsonify({
        "message": f"Kontener '{container_name}' został zrestartowany"
    }), 200

The restart operation is asynchronous. The API returns after initiating the restart, but the container may take a few seconds to fully restart.

Delete Container

Delete a container instance and release all resources.

Endpoint

DELETE /api/azure/containers

Request Body

{
  "subscriptionId": "12345678-1234-1234-1234-123456789abc",
  "resourceGroup": "my-rg",
  "containerName": "my-container"
}

Parameters

subscriptionId

string

required

Azure subscription ID

resourceGroup

string

required

Resource group name

containerName

string

required

Container group name to delete

Response

{
  "message": "Kontener 'my-container' został usunięty"
}

Implementation

def delete_container():
    data = request.get_json()
    subscription_id = data.get("subscriptionId")
    resource_group = data.get("resourceGroup")
    container_name = data.get("containerName")
    
    credential = FlaskCredential()
    client = ContainerInstanceManagementClient(credential, subscription_id)
    
    client.container_groups.begin_delete(
        resource_group,
        container_name
    ).wait()
    
    return jsonify({
        "message": f"Kontener '{container_name}' został usunięty"
    }), 200

Deleting a container is permanent and cannot be undone. Ensure you have backups of any important data before deletion.

Error Handling

Common error responses:

Unauthorized (401)

{
  "error": "Unauthorized"
}

The access token is missing or invalid. Re-authenticate to obtain a valid token.

Bad Request (400)

{
  "error": "Brak wymaganych danych"
}

Required parameters are missing from the request.

Server Error (500)

{
  "error": "The subscription is not registered to use namespace 'Microsoft.ContainerInstance'"
}

The Azure subscription does not have the Container Instances resource provider registered.

Best Practices

Resource Sizing

Start with minimal resources (0.5 CPU, 0.5 GB memory) and scale up as needed
Monitor CPU and memory usage to right-size containers
Consider Azure Container Apps for production workloads requiring auto-scaling

Image Management

Use specific image tags instead of latest for production
Store images in Azure Container Registry for better performance
Enable vulnerability scanning in ACR
Use managed identity instead of registry passwords when possible

Security

Store registry credentials in Azure Key Vault
Use private endpoints for ACR access
Implement network security groups for container traffic
Regularly update base images to patch vulnerabilities

Cost Optimization

Delete unused containers to avoid charges
Use restart policy OnFailure or Never for batch jobs
Consider Azure Container Apps for better price/performance
Use spot instances for fault-tolerant workloads

Container Monitoring

Monitor Azure Container Instances with metrics and logs

Container Alerts

Configure alerts for container health and performance

Get Started

Authentication

Core Features

Virtual Machines

Containers

Storage

Networking

Monitoring & Logs

Deployment

​Azure Container Instances

​List Containers

​Endpoint

​Authentication

​Response

​Implementation

​Create Container

​Endpoint

​Request Body

​Parameters

​Response

​Implementation

​Container Configuration Details

​Restart Container

​Endpoint

​Request Body

​Parameters

​Response

​Implementation

​Delete Container

​Endpoint

​Request Body

​Parameters

​Response

​Implementation

​Error Handling

​Unauthorized (401)

​Bad Request (400)

​Server Error (500)

​Best Practices

​Resource Sizing

​Image Management

​Security

​Cost Optimization

​Related Documentation

Container Monitoring

Container Alerts

Build docs developers (and LLMs) love

Azure Container Instances

List Containers

Endpoint

Authentication

Response

Implementation

Create Container

Endpoint

Request Body

Parameters

Response

Implementation

Container Configuration Details

Restart Container

Endpoint

Request Body

Parameters

Response

Implementation

Delete Container

Endpoint

Request Body

Parameters

Response

Implementation

Error Handling

Unauthorized (401)

Bad Request (400)

Server Error (500)

Best Practices

Resource Sizing

Image Management

Security

Cost Optimization

Related Documentation