|
| 1 | +# Run our Flask app with gunicorn in Docker |
| 2 | + |
| 3 | +Throughout the course, we've been working with a Docker image like this one: |
| 4 | + |
| 5 | +```dockerfile |
| 6 | +FROM python:3.10 |
| 7 | +EXPOSE 5000 |
| 8 | +WORKDIR /app |
| 9 | +COPY requirements.txt . |
| 10 | +RUN pip install -r requirements.txt |
| 11 | +COPY . . |
| 12 | +CMD ["flask", "run", "--host", "0.0.0.0"] |
| 13 | +``` |
| 14 | + |
| 15 | +This is all well and good for local development, but when we deploy our application we want to run it with the best performance possible. |
| 16 | + |
| 17 | +This is why we don't want to run the Flask development server and the Flask debugger. Instead, we'll use gunicorn to run our app. |
| 18 | + |
| 19 | +## Run our Flask app with gunicorn |
| 20 | + |
| 21 | +First let's add `gunicorn` to our `requirements.txt` file: |
| 22 | + |
| 23 | +```text title="requirements.txt" |
| 24 | +flask |
| 25 | +flask-smorest |
| 26 | +python-dotenv |
| 27 | +sqlalchemy |
| 28 | +flask-sqlalchemy |
| 29 | +flask-jwt-extended |
| 30 | +passlib |
| 31 | +flask-migrate |
| 32 | +# highlight-start |
| 33 | +gunicorn |
| 34 | +# highlight-end |
| 35 | +``` |
| 36 | + |
| 37 | +Then, let's change our `Dockerfile` to use `gunicorn`: |
| 38 | + |
| 39 | +```dockerfile |
| 40 | +FROM python:3.10 |
| 41 | +WORKDIR /app |
| 42 | +COPY ./requirements.txt requirements.txt |
| 43 | +# highlight-start |
| 44 | +RUN pip install --no-cache-dir --upgrade -r requirements.txt |
| 45 | +# highlight-end |
| 46 | +COPY . . |
| 47 | +# highlight-start |
| 48 | +CMD ["gunicorn", "--bind", "0.0.0.0:80", "app:create_app()"] |
| 49 | +# highlight-end |
| 50 | +``` |
| 51 | + |
| 52 | +The `CMD` line change is the important one, as it runs `gunicorn` on port `80`, and we pass in the app factory function. |
| 53 | + |
| 54 | +:::tip |
| 55 | +Note I've also changed the `pip install` line. Adding `--no-cache-dir` and `--upgrade` just makes sure we can't accidentally install from a cache directory (which shouldn't exist anyway!), and that we'll upgrade to the latest possible versions allowed by our `requirements.txt` file. |
| 56 | +::: |
| 57 | + |
| 58 | +## Run the Docker container locally with the Flask development server and debugger |
| 59 | + |
| 60 | +If you use this `Dockerfile`, it doesn't mean you can't run it locally using the Flask development server. You don't have to lose the automatic restarting capabilities, or the Flask debugger. |
| 61 | + |
| 62 | +To run the Docker container locally, you'll have to do this from now on: |
| 63 | + |
| 64 | +```zsh |
| 65 | +docker run -dp 5000:5000 -w /app -v "$(pwd):/app" teclado-site-flask sh -c "flask run" |
| 66 | +``` |
| 67 | + |
| 68 | +This is similar to how we've ran the Docker container with our local code as a volume (that's what `-w /app -v "$(pwd):/app"` does), but at the end of the command we're telling the container to run `flask run` instead of the `CMD` line of the `Dockerfile`. That's what `sh -c "flask run"` does! |
| 69 | + |
| 70 | +Now you're ready to commit and push this to your repository and re-deploy to Render.com! |
0 commit comments