This is an Ansible playbook template accompanying flask-minimal cookiecutter template. It provides means to deploy flask-minimal application on an Ubuntu server.

Install cookiecutter and create a Flask application as per flask-minimal and build it (make sdist).

Create accompanying Ansible project from this template. Put it side-by-side with the Flask application created above (in the same root directory), and use same value for application package name, or you will need to change generated configuration accordingly:

cookiecutter https://github.com/candidtim/cookiecutter-flask-ansible.git

Install prerequisites as per generated README, and then run the application in the Virtualbox to test the generated configuration:

cd yourapplication-ansible
vagrant up

And open deployed application at http://127.0.0.1:8080/

When ready, configure inventory file and deploy to the server with:

ansible-playbook playbook.yml -i inventory.ini

• setuptools-ready Flask application installs into a dedicated virtualenv
• runs in a uWSGI web server instance, installed into the same virtualenv
• nginx server is used as a reverse proxy
• all this is installed on an Ubuntu server, with a systemd service to run uWSGI instance
• for tests only, Vagrant and Virtualbox are used

This implementation also assumes an Ubuntu development machine.

This template is much more opinionated compared to the flask-minimal, but it still strives to be minimal and production-ready in its own way.

Deploying Flask application to production implies a lot of choice: OS version, WSGI implementation, reverse proxy HTTP server, logging configuration, maybe other ones. It is hard to provide a solution flexible enough to fit all possible needs: such an attempt will likely result in a very complex solution, and will still require customizations in most of real-world scenarios. More importantly, it is rarely required: for example, most projects won't require to be deployed to different WSGI servers at the same time, but stick to one of them.

This is why this project, instead, provides a simple solution, which is ready for use, but makes a couple of choices. You can use this in two cases:

• if choices made here fit your needs
• or as an example to build your own configuration

... are welcome! Feel free to create a pull request to fix bugs, keep up to date, or propose a feature.