AI-driven Application Containers on Azure

Question

Pulumi · Accepted Answer

To deploy AI-driven application containers on Azure, you would typically use Azure Container Apps, a service that allows you to run containers in a serverless environment. This service is suitable for microservices architecture and provides dynamic scaling, including scaling to zero when the application isn't in use.

In the program below, an Azure Container App will be set up, which could host any container capable of running an AI-driven application. This program will define the infrastructure necessary to deploy a container, including setting up an environment for the container app and configuring the app itself with the desired container image.

To start, you'll need the following setup on your local development environment:

Python 3
Pulumi CLI
Azure CLI

Here's the step-by-step Pulumi program in Python that sets up an AI-driven application container on Azure:

Define the Azure Container App environment within your resource group. A container app environment is a secured and dedicated portion of Azure where you will deploy your container apps.
Deploy a container app within the environment you just created. This container app will run your AI-driven application, which is assumed to be containerized.

Here is the Pulumi program to do this:

import pulumi
import pulumi_azure_native as azure_native

# Create an Azure Resource Group
resource_group = azure_native.resources.ResourceGroup("resource_group")

# Create an Azure Container Apps Environment
container_app_environment = azure_native.app.ContainerAppEnvironment("containerAppEnvironment",
    resource_group_name=resource_group.name,
    location=resource_group.location,
    configuration=azure_native.app.ContainerAppEnvironmentConfigurationArgs(
        app_logs_configuration=azure_native.app.AppLogsConfigurationArgs(
            destination=azure_native.app.LogDestinationArgs(
                name="log-analytics",
                log_analytics=azure_native.app.LogAnalyticsArgs(
                    workspace_id="<Your-Log-Analytics-Workspace-ID>",
                    workspace_key="<Your-Log-Analytics-Workspace-Key>",
                    # Fill in these details with your Azure Log Analytics Workspace ID and Key to enable monitoring and logging
                ),
            ),
        ),
    ),
)

# Create an Azure Container App
container_app = azure_native.app.ContainerApp("containerApp",
    resource_group_name=resource_group.name,
    location=resource_group.location,
    managed_environment_id=container_app_environment.id,
    configuration=azure_native.app.ContainerAppConfigurationArgs(
        ingress=azure_native.app.IngressArgs(
            external=True,
            target_port=80
            # If your container serves HTTP traffic, ensure it listens on port 80, or change this to the port your app listens on.
        ),
        secrets=[azure_native.app.SecretArgs(name="example-secret", value="s3cr3t")],
        registries=[azure_native.app.RegistryCredentialsArgs(
            server="mycontainerregistry.azurecr.io",
            username="myregistryusername",
            password_secret_ref="example-secret",
        )]
    ),
    template=azure_native.app.TemplateArgs(
        containers=[
            azure_native.app.ContainerArgs(
                name="ai-app-container",
                image="mycontainerregistry.azurecr.io/ai-application:latest",
                # Replace with your container image that runs the AI application.
                resources=azure_native.app.ContainerResourcesArgs(
                    cpu=0.5,
                    memory="1.0Gi"
                    # Specify the amount of CPU and memory that your application needs.
                ),
            ),
        ],
    ),
)

# Export the URL of the Container App
pulumi.export("container_app_url", container_app.configuration.apply(lambda c: c.ingress.fqdn))

In this program, replace the mycontainerregistry.azurecr.io/ai-application:latest with the URL of the container image you'd like to deploy. It's also important to configure the workspace_id and workspace_key with your Azure Log Analytics Workspace ID and Key to enable monitoring and logging properly.

This demonstrates how you can deploy an application on Azure Container Apps through Pulumi. The actual AI application logic needs to be packaged in a container and pushed to a registry that your Pulumi script references in the image parameter.

Make sure your application listens on the port you configure in the target_port of the ingress configuration. If you want to use another port rather than 80 make the necessary adjustment.

Remember to have the necessary Azure permissions and configurations set up before running your Pulumi stack, including authentication to the Azure API and access rights to create the various resources defined in your code.