Azure Streamlit Deployment

Dummy project to enable the deployment of Streamlit Apps on Microsoft Azure.

Table of Contents

• Introduction
• Prerequisites
• Deployment
• Continuous Integration and Deployment (CICD)
  • Azure Resources
  • Docker
  • Streamlit App on Azure
• Challenges)

Introduction

The purpose of this project is to create a dummy project set up, with guidance and steps, to enable the delopyment of Streamlit Apps on Microsoft Azure. Something that took me a bit of time to crack!

Prerequisites

Azure Subscription with the ability to create resources, Docker, VS Code Azure Extension (optional), VS Code Docker Extension (optional), Python ~3.9

Deployment

To deploy your Streamlit app on Azure, use Docker to containerize the app, Azure Container Registry (ACR) to store the Docker image, and an App Service Plan to host and run the app on Azure.

CI/CD

To provide automated deployment of the Streamlit app, a Continuous Integration and Deployment (CICD) pipeline can be set up with branch policies. This pipeline will leverage Azure DevOps to automatically build and push a new container image to the Azure Container Registry (ACR) and subsequently the Azure App Service whenever there's a new commit to the main branch. See build-pipeline.yml.

Azure Resources

Creating Azure Resources

Before deploying the app, the necessery Azure resources need to created This can be done using either Azure CLI, PowerShell, Azure Portal UI or Visual Studio Code (VS Code) extensions. I opted for the VS Code extension route.

Resources include: Azure Container Registery and Azure App Service Plan (note this can be created later at the time of pushing the containerised image to the app)

Docker

Building and Pushing the Docker Image

Now that we have our Azure resources in place, let's build the Docker image for the Streamlit app and push it to the Azure Container Registry (ACR).

1. Make sure you have Docker installed on your local machine.

2. In the root directory of your Streamlit app project in your IDE, create a Dockerfile. This file will define the environment and dependencies required to run your app.

3. Build the Docker image - this can be done most easily via the VS Code Docker Extension, once installed, just right click on the Dockerfile and click 'Build Image'. This can be done locally and sent to Azure later or built directly within the Azure Container Registry. Alternatively, the command line can be used to build the docker image and push to Azure...

docker build -t your-acr-name.azurecr.io/your-app-name:latest .

4. Login to Azure Container Registry:

az acr login --name YourACRName

5. Push the Docker image to Azure Container Registry:

docker push your-acr-name.azurecr.io/your-app-name:latest

Streamlit App on Azure

Deploying the App to App Service

With the Docker image pushed to the Container Registry, the Streamlit App can be deployed to an Azure App Service using the previously created App Service Plan.

Deploy via VC code

1. Ensure the Docker extension is installed on VS Code

2. Click on "Registries", then 'Connect Registry, then 'Azure'

3. Once connected you should see the Azure Container Registry created earlier.

4. You can right click on the containerised image within the new registry and click 'Deploy Image to Azure App Service' to deploy the image to the new application

5. Access your deployed Streamlit app using the App Service URL.

Deploy via Azure Portal UI

1. Go to the Azure Portal and navigate to your App Service.

2. Click on "Deployment Centre" to set up continuous deployment using your Git repository or other version control systems for automatic deployment whenever there's a new commit to the main branch.

3. In the "Configuration" tab of your App Service, set the "Docker Container" as the deployment method and provide the URL of your Docker image from the Azure Container Registry.

4. Save the configuration, and Azure will automatically deploy your Streamlit app from the Docker image.

5. Access your deployed Streamlit app using the App Service URL.

Deploy via CLI

• to be added

Challenges

In your Dockerfile, make sure to expose the port on which your Streamlit app runs (typically and by default port 8501 for Steamlit) using the EXPOSE command