django-pubsubpull is intended to help you decouple micro-services.
It provides mechanisms that allow services to publish changes, subscribe
to change notifications, and pull records.
In practice what we’ll end up with is along the lines of:
- Use Django slumber and async to provide a load of basic
infrastructure this is going to need.
- Provide new models for tracking database level changes to certain
tables (models), and database triggers to ensure that they are
- All operations will be based on HTTP using GET, PUT and DELETE only.
- Django middleware to capture as much information about the request
and the underlying changes we can and to tie that in to the records
of database changes recorded there.
- Support for Django 1.0 through 1.7. We may support 1.8 and later if
we can make our middleware subsume the functionality of the old
- Only support Postgres 9.4 and later.
The use cases we envisage include:
- A service needs to locally cache a copy of some attributes of an
object held in another micro-service, generally for performance
reasons (for example, a billing service needing copies of a
customer’s location so that it can choose the correct rates).
- A service needs to record meta-data about user instructed changes for
audit trail purposes (for example, any change to a customer email
subscription is recorded with full details of who made the change and
when it was made).
- A service needs to be notified of certain API calls so that it can in
turn trigger other behaviours (for example, a comms service sending a
welcome email when a new user is created).
The three parts are:
- Publish: A service will notify it’s subscribers about any changes to
the requested model instances. Ultimately the publish will hook into
a database trigger to ensure that changes are always properly
recorded no matter how it occurs.
- Subscribe: A service can request change notifications be sent to it
from the publishing service. The subscriber is able to specify a URL
that is used for a PUT request when changes are notified.
- Pull: A service can iterate over all of the instances of a particular
model and then periodically iterate over new ones.
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.