Usage

Using eventlog is pretty simple. Throughout your site, you just call a single function, log() to record whatever information you want to log. If you are wanting to log things from third party apps, your best bet is to use signals. Hopefully the app in question provides some useful signals, but if not, perhaps some of the built in model signals will be enough (e.g. pre_save, post_delete, etc.)

Example:

from eventlog.models import log

def some_view(request):
    # stuff is done in body of view
    # then at the end before returning the response:
    log(
        user=request.user,
        action="CREATED_FOO_WIDGET",
        extra={
            "pk": foo.pk,
            "title": foo.title
        }
    )
    return HttpResponse()

The action parameter can be any string you choose. By convention, we always use all caps. Take note, however, whatever you choose, will be the label that appears in the admin’s list filter, so give it some thought on naming conventions in your site so that the admin interface makes sense when you have 50,000 log records you want to filter down and analyze.

The extra parameter can be anything that will serialize to JSON. Results become easier to manager if you keep it at a single level. Also, keep in mind that this is displayed in the admin’s list view so if you put too much it can take up a lot of space. A good rule of thumb here is put enough identifying data to get a sense for what is going on and a key or keys that enable you to dig deeper if you want or need to.

Mixin

You can also easily make your class based views auto-logged by using the eventlog.mixins.EventLogMixin. The only requirement is defining an action_kind property on the view. But you can also override a number of properties to customize what is logged.