Django Message Broker is a plugin written in Python for Django that provides an all-in-one message broker. It interfaces with Django Channels and Celery [1], and replaces the need for separate message brokers such as Redis and RabbitMQ.
The motivating use case for Django Message Broker is small site solutions where it is easier to deploy a single server containing all the required functionality rather than a multi-server solution. An example would be a small site running data science models, where executing the model is a long running process and needs to execute in the background so that it doesn’t degrade the responsiveness of the user interface.
Potential scenarios for Django Message Broker include:
- Prototyping, Testing, Training
- Data science projects where the model complexity exceeds the capabilities of packages such as Shiny, Dash and Streamlit.
- Small systems with a low number of users.
The Django Message Broker is an installable component of the Django platform that provides an easy to install, all-in-one alternative for small scale solutions. It does not replace the need for high volume message brokers where message volume and redundancy are important.
Note
- The Celery interface is in development and not supported in this release.
Install latest stable version into your python environment using pip::
pip install django-message-broker
Once installed add django_message_broker to your INSTALLED_APPS in settings.py:
INSTALLED_APPS = (
'django_message_broker',
...
)
Django Message Broker should be installed as early as possible in the installed applications list and ideally before other applications such as Channels and Celery. The message broker forks a background process which should occur before other applications create new threads in the current process.
To configure the Django Message Broker as a Channels layer add the following to the CHANNEL_LAYERS
setting in settings.py:
CHANNEL_LAYERS = {
'default': {
'BACKEND': 'django_message_broker.ChannelsServerLayer',
},
}