Django on Kubernetes

Django on Kubernetes

This is part 2 of our Kubernetes hackday series. Part 1 goes over how we spent the day and what the goals and motivations are.

For part 2, we're going to delve into the architecture required for running a Django application on Kubernetes, as well as some of the tooling we used to assist us.
This post will assume some knowledge with deploying and operating a production web application. I'm not going to spend much time going over the terminology that Kubernetes uses either. What I hope to do in this post is present enough information to kickstart your own migration. Kubernetes is big - and knowing what to research now and what to put off until later is really tricky.
I'm also going to describe this deployment in terms of Amazon EKS simply because I was on the EKS team, and most of our applications are already on AWS.

Faster Django Tests by Disabling Signals

Django signals are a form of Inversion Of Control that developers can use to trigger changes based (usually) on updates to models.

The canonical example is automatically creating a UserProfile when a User instance is created:

from django.conf import settings
from django.db.models.signals import post_save
from django.dispatch import receiver

@receiver(post_save, sender=settings.AUTH_USER_MODEL)
def post_save_receiver(sender, instance, created, **kwargs):
    UserProfile.objects.create(user=instance, ...)

Signals are useful for connecting your own behaviour to events that you might not have control over, such as those generated by the framework or libary code. In general, you should prefer other methods like overriding model save methods if you have the ability, as it makes code easier to reason about.

Signal receivers can sometimes be much more complex than creating a single object, and there can be many receivers for any given event, which multiply the time it takes to perform simple actions.

Tests and signals

Often you won't actually need your signals to execute when you're running your test suite, especially if you're creating and deleting many thousands of model instances. Disconnecting signals is tricky though, especially if the disconnect/reconnect logic can be stacked.

An easier, but more invasive method for suspending signal receivers is to check a global setting, and return early.

from django.conf import settings

@receiver(post_save, sender=MyModel)
def mymodel_post_save(sender, **kwargs):
    if settings.DEBUG:

This has two drawbacks. Firstly, it's messy, and you need to remember to add the check to each receiver. Secondly, it depends on a specific setting that you might need to have turned off when running your tests, which makes it harder to test the actual receiver.

Selectively disabling signals

We can take this most recent idea of checking a setting, and wrap it up nicely in our own decorator. When running your tests, you can override the SUSPEND_SIGNALS setting per test method or class.

Here's a gist that does just that

import functools

from django.conf import settings
from django.dispatch import receiver

def suspendingreceiver(signal, **decorator_kwargs):
    def our_wrapper(func):
        @receiver(signal, **decorator_kwargs)
        def fake_receiver(sender, **kwargs):
            if settings.SUSPEND_SIGNALS:
            return func(sender, **kwargs)
        return fake_receiver
    return our_wrapper

And using this decorator in your test suite is straightforward:

class MyTestCase(TestCase):
    def test_method(self):
        Model.objects.create()  # post_save_receiver won't execute

@suspendingreceiver(post_save, sender=MyModel)
def mymodel_post_save(sender, **kwargs):

And just like that, we can skip signal receivers in our test suite as we like!

Kogame (Koh-Gah-Mi) - A real time game in Django

Kogame (Koh-Gah-Mi) - A real time game in Django

For our March hackday this year we decided to build a multiplayer game using

[Django Channels]( We've been keeping

an eye on channels over the time, and thought with the release of channels 2.0 that

it was the right time to dive in and get some practical experience. The organisers

didn't want to do yet-another-chat-system implementation, so they decided to make

things a bit more interesting, and look at writing a real-time game.