Under Construction

heroku-django-docker-example

The smallest "reasonable" project for running a Django project under Heroku Docker

There are a few examples of how to run Python programs in Docker, but there is not a complete example that brings together a lot of practical considerations. This is my attempt to rectify that.

I have tried to add comments liberally throughout the code.

Project Overview

When building a Python project in Heroku, Heroku has some (strong) suggestions about how to handle different use-cases.

• Security credentials in environment variables
• Django for web-serving
• dj_database_url for converting the DATABASE_URL to a Django database configuration
• Django-RQ for dealing with background tasks
• WhiteNoise for serving static assets (images, CSS, etc)

I will try to illustrate all of those.

In addition, you probably found this example because you were trying to figure out how to use Numpy, Scipy, Pandas, etc with Heroku. So I'll add that just for fun.

Note: This app depends on Redis (for Django-RQ) and Postgres, and I'm assuming that you are running them as "normal" processes on localhost, and not trying to run them within a Docker container.

Docker Setup and Use

Dockerfile Structure

We have a base Dockerfile that prepares an image for us. Then we have a series of related Dockerfiles (Dockerfile.web, Dockerfile.larry, Dockerfile.curly) that inherit from the base Dockerfile.

Building Docker Images

Base Image

To build the base image, do:

$ docker build --tag example --file Dockerfile .

Note that the use of "example" as a tag is important -- it is referenced in the "FROM" lines of the other Dockerfiles

Other Images

To build a specific image:

$ docker build --tag example/web --file Dockerfile.web .

(replace "web" with the specfic file you want)

If I Change "X", What Do I Have to Rebuild?

• If you add something to the operating system (e.g. "apt-get install"), add that to your Dockerfile (near line 8), and rebuild the base image
• If you add something to your virtualenv (e.g. "pip install"), add that to your requirements.txt, and rebuild the base image
• If you change Python code, rebuild the image you are working on

Running Individual Containers

The Web Process

You have to specify the ports to get everything tied together

$ docker run -p 8000:8000 -e PORT=8000 example/web

Worker Processes

$ docker run example/larry  (or curly, or moe)

Running Everything with Docker-Compose

Being able to run your containers individually is great, but you don't want to have to start and stop services individually. That's where docker-compose comes in.

The supplied docker-compose.yml file will allow you to start all your services with

$ docker-compose up --build

Note that if you change something that requires the base image to be re-built (adding to your Unix installation with apt-getor your Python environment with pip), you have to rebuild the base image yourself, using

$ docker build --tag=example --file=Dockerfile .

Environment Variables

Eventually, you may end up with a situation where you will need environment variables at build time (usually because they are referenced in your settings file). If that is necessary, you can safely put these in your Dockerfile.web as

ENV MYVARIABLE placeholder

And Heroku will override that "placeholder" value at runtime.

Deploying To Heroku

Once you have things set up and are ready to push to Heroku.

$ docker build --tag=example --file=Dockerfile
$ heroku container:push --recursive --app=name-of-your-app
$ heroku container:release --app=name-of-your-app

And that is it!