Migrating static content into containers was a great way to learn the basics of Docker, but real-world uses are usually dynamic, including full applications. This lab will show you the process to dockerize a Flask application. Flask is a lightweight Python WSGI micro web framework, however, you won’t need to know any Python to complete this lab.
Learning Objectives
Successfully complete this lab by achieving the following learning objectives:
- Create the Build Files
- Inspect the Flask application files to learn what files we need to include and exclude from the image. There is a
Pipfile.lock, a.gitignore, and amigrationsdirectory. We don’t want those in the image. - Create a
.dockerignorefile to exclude build and metadata information from the image. - Create the
Dockerfileto build the image.- Start with Python 3 as the base image.
- Setup Python environment variables.
- Install dependencies in the container from the Pipfile.
- Set the working directory.
- Specify the port flask will listen on.
- Start the flask server.
- Inspect the Flask application files to learn what files we need to include and exclude from the image. There is a
- Build and Setup Environment
- Build an image named
notesappwith version0.1. - Inspect the Docker environment to see the new image, and what else is running.
- Use a container of the notesapp image to set up the database.
Note: The database can be set up via the app itself using the following command.
flask db init && flask db migrate && flask db upgrade
- Build an image named
- Run, Evaluate, and Upgrade
- Run the NotesApp container in the current terminal to watch its output.
- View the application in a browser. Sign up for an account, create a note, then edit it.
- View the messages in the terminal and check for warnings.
- Stop using Flask in debug mode.
- Remove the line
FLASK_ENV=developmentfrom the file~/notes/.env.
- Remove the line
- Since we’ve changed a file included in the image, build the image again as version
0.2. - Run the new version of the NotesApp image in the terminal. Debug mode should no longer be on.
- Look for any new warnings.
Note: We’re not in debug anymore, but now Flask tells us "Do not use it in a production deployment". That’s not a good way to leave our container.
- Stop the container.
- Upgrade to Gunicorn
We will upgrade our Flask application to use Gunicorn, which is a production-ready WSGI HTTP server. This will involve adding a bit of code to our existing app and upgrading our build.
- Add Gunicorn to the dependencies using pipenv.
- Pipenv is being used to manage the dependencies for our application. Let’s use the image we have, which includes
pipenv, to upgrade the Pipfile so we don’t have to install a development environment on the server.
- Pipenv is being used to manage the dependencies for our application. Let’s use the image we have, which includes
- Upgrade the application code for Gunicorn.
- Flask can use
dotenvautomatically, but Gunicorn must be told to use it explicitly. Add the following lines to__init__.pybelow the import statements:from dotenv import load_dotenv, find_dotenv load_dotenv(find_dotenv())
- Flask can use
- Change the Dockerfile to run Gunicorn instead of Flask.
- Replace the
WORKDIRandCMDin the Dockerfile with the below code:WORKDIR /app CMD [ "gunicorn", "-b 0.0.0.0:80", "notes:create_app()"]
- Replace the
- Add Gunicorn to the dependencies using pipenv.
- Build a Production Image
- Build the NotesApp image again. Remember to increment the version.
- Run a container from the production image, this time using detached mode to run in the background.
- Inspect the Docker environment to ensure everything is running.
- View the application in a browser. Use the server’s public IP.
