This isn't the best fix because we're still using NODE_ENV=development
in a "production" environment. Ideally we would stop doing this.
Right now, the Helm charts are broken because they deploy the backend
straight into Error state.. This change will unbreak them for the
moment.
Add a Dockerfile which can be used to build an image that the helm
charts will use as default. These images require mkdocs and cookiecutter
to be installed (for now) in order to work around the docker requirement
in kubernetes.