Awesome

Butler

Butler is for automating Python app deployments to AWS EC2. Currently it installs everything on one machine. As the build process lasts about a minute it makes scaling up and down very simple. It tries to be as easy as Heroku but you are in full control.

Quickstart

1. Download this package to your disk.
2. cd to the download directory
3. Change the env.hosts and env.key_filename to reflect your own
4. fab build

This builds everything and you should soon have a working deployment.

Default stack

• NGINX
• uWSGI
• MySQL
• Memcached

TODO

• Apache installation and configuration
• Gunicorn installatiin and configuration
• Postgres installation and configuration
• Varnish installation and configuration

• AWS instance spawning automation with Boto

• Currently there is a default stack but this will change in the near future when you'll have the option of choosing your stack.
• Make it possible to deploy the DB and the application of separate machines with optional replication.
• Possibility of running a load balancer in front of your web servers. (HAproxy)
• Background workers with celery or redis queues

• Build a optional UI for deployment and monitoring with stats from CloudWatch

Settings

For a django application you could use settings.py to access or load settings.

from fabric.contrib import django
django.settings_module('myproject.settings')
from django.conf import settings

See http://readthedocs.org/docs/fabric/en/latest/api/contrib/django.html

Requirements

• Fabric - https://github.com/fabric/fabric by Christian Vest Hansen and Jeffrey E. Forcier
• git - http://git-scm.com/
• Boto - https://github.com/boto/boto by Mitch Garnaat