A lot of Django app deployments over Amazon’s EC2 use HTTP servers NGINX and Gunicorn.
I was wondering what they actually do and why both are used in parallel. What is the purpose of running them both in parallel?
They aren’t used in parallel. NGINX is a reverse proxy. It’s first in line. It accepts incoming connections and decides where they should go next. It also (usually) serves static media such as CSS, JS and images. It can also do other things such as encryption via SSL, caching etc.
Gunicorn is the next layer and is an application server. NGINX sees that the incoming connection is for
www.domain.com and knows (via configuration files) that it should pass that connection onto Gunicorn. Gunicorn is a WSGI server which is basically a:
simple and universal interface between web servers and web applications or frameworks
Gunicorn’s job is to manage and run the Django instance(s) (similar to using
django-admin runserver during development)
The contrast to this setup is to use Apache with the
mod_wsgi module. In this situation, the application server is actually a part of Apache, running as a module.