How To Create Minimal Docker Images for Python Applications - KDNuggets (2024)

This tutorial will teach you how to create minimal Docker images for Python applications.

By Bala Priya C, KDnuggets Contributing Editor & Technical Content Specialist on June 25, 2024 in Data Engineering

Image by Editor | Midjourney & Canva

Creating minimal Docker images for Python apps enhances security by reducing the attack surface, facilitates faster image builds, and improves overall application maintainability. Let’s learn how to create minimal Docker images for Python applications.

Prerequisites

Before you get started:

• You should have Docker installed. Get Docker for your operating system if you haven’t already.
• A sample Python application you need to build the minimal image for. You can also follow along with the example app we create.

Create a Sample Python Application

Let's create a simple Flask application for inventory management. This application will allow you to add, view, update, and delete inventory items. We'll then dockerize the application using the standard Python 3.11 image.

In your project directory, you should have app.py, requirements.txt, and Dockerfile:

inventory_app/├── app.py├── Dockerfile├── requirements.txt

Here’s the code for the Flask app for inventory management:

# app.pyfrom flask import Flask, request, jsonifyapp = Flask(__name__)# In-memory database for simplicityinventory = {}@app.route('/inventory', methods=['POST'])def add_item(): item = request.get_json() item_id = item.get('id') if not item_id: return jsonify({"error": "Item ID is required"}), 400 if item_id in inventory: return jsonify({"error": "Item already exists"}), 400 inventory[item_id] = item return jsonify(item), 201@app.route('/inventory/', methods=['GET'])def get_item(item_id): item = inventory.get(item_id) if not item: return jsonify({"error": "Item not found"}), 404 return jsonify(item)@app.route('/inventory/', methods=['PUT'])def update_item(item_id): if item_id not in inventory: return jsonify({"error": "Item not found"}), 404 updated_item = request.get_json() inventory[item_id] = updated_item return jsonify(updated_item)@app.route('/inventory/', methods=['DELETE'])def delete_item(item_id): if item_id not in inventory: return jsonify({"error": "Item not found"}), 404 del inventory[item_id] return '', 204if __name__ == '__main__':app.run(host='0.0.0.0', port=5000)

This is a minimal Flask application that implements basic CRUD (Create, Read, Update, Delete) operations for an in-memory inventory database. It uses Flask to create a web server that listens for HTTP requests on port 5000. When a request is received:

• For a POST request to /inventory, it adds a new item to the inventory.
• For a GET request to /inventory/<item_id>, it retrieves the item with the specified ID from the inventory.
• For a PUT request to /inventory/<item_id>, it updates the item with the specified ID in the inventory.
• For a DELETE request to /inventory/<item_id>, it deletes the item with the specified ID from the inventory.

Now create the requirements.txt file:

Flask==3.0.3

Next create the Dockerfile:

# Use the official Python 3.11 imageFROM python:3.11# Set the working directoryWORKDIR /app# Install dependenciesCOPY requirements.txt requirements.txtRUN pip install --no-cache-dir -r requirements.txt# Copy the current directory contents into the container at /appCOPY . .# Expose the port the app runs onEXPOSE 5000# Run the applicationCMD ["python3", "app.py"]

Finally build the image (we use the tag full to identify that this uses the default Python image):

$ docker build -t inventory-app:full .

Once the build is complete you can run the docker images command:

$ docker imagesREPOSITORY TAG IMAGE ID CREATED SIZEinventory-app full 4e623743f556 2 hours ago 1.02GB

You’ll see that this super simple app is about 1.02 GB in size. Well, this is because the base image we used the default Python 3.11 image has a large number of Debian packages and is about 1.01 GB in size. So we need to find a smaller base image.

Well, here are the options:

• python:version-alpine images are based on Alpine Linux and will give you the smallest final image. But you need to be able to install packages as well, yes? But that’s a challenge with alpine images.
• python:version-slim comes with the minimal number of Debian packages needed to run Python. And you’ll (almost always) be able to install most required Python packages with pip.

So your base image should be small. But not too small that you face compatibility issues and wrap your head around installing dependencies (quite common for Python applications). That’s why we’ll use the python:3.11-slim base image in the next step and build our image.

Choosing the Optimal Base Image | Image by Author

Use the Slim Python Base Image

Now rewrite the Dockerfile to use the python:3.11-slim base image like so:

# Use the official lightweight Python 3.11-slim imageFROM python:3.11-slim# Set the working directoryWORKDIR /app# Install dependenciesCOPY requirements.txt requirements.txtRUN pip install --no-cache-dir -r requirements.txt# Copy the current directory contents into the container at /appCOPY . .# Expose the port the app runs onEXPOSE 5000# Run the applicationCMD ["python3", "app.py"]

Let’s build the image (tagged slim):

$ docker build -t inventory-app:slim .

The python:3.11-slim base image is of size 131 MB. And the inventory-app:slim image is around 146 MB which is much smaller than the 1.02GB image we had earlier:

$ docker imagesREPOSITORY TAG IMAGE ID CREATED SIZEinventory-app slim 32784c60a992 About an hour ago 146MBinventory-app full 4e623743f556 2 hours ago 1.02GB

You can also use multi-stage builds to make the final image smaller. But that's for another tutorial!

Additional Resources

Here are a few useful resources:

• Containerize Python Apps with Docker in 5 Easy Steps
• python - Official Image | Docker Hub
• Differences Between Standard Docker Images and Alpine \ Slim Versions

Bala Priya C is a developer and technical writer from India. She likes working at the intersection of math, programming, data science, and content creation. Her areas of interest and expertise include DevOps, data science, and natural language processing. She enjoys reading, writing, coding, and coffee! Currently, she's working on learning and sharing her knowledge with the developer community by authoring tutorials, how-to guides, opinion pieces, and more. Bala also creates engaging resource overviews and coding tutorials.