django-performance-metrics

This library helps you measure and record the performance of functions within your complex Django applications. It looks a bit like this:

from django.shortcuts import render
from metrics import profile, measure
import time

@profile()
def expensive_sub_function():
    time.sleep(3)
    measure("some_value", 150)
    measure("some_other_value", 1337)

@profile()
def do_something_expensive():
    time.sleep(1)
    expensive_sub_function()

def my_view(request):
    do_something_expensive()
    return render(request, "template.html")

After adding metrics to your INSTALLED_APPS, metrics.middleware.MetricsMiddleware to your MIDDLEWARE_CLASSES you get the following output in the console:

 - response.time: 4019.8562145233154
   * Meta: is_authenticated=False username=None is_staff=False referrer=None is_superuser=False view=app.views.test_long is_ajax=False method=GET url=/ module=app.views
	 - app.views.do_something_expensive: 4.002618789672852
		 - app.views.expensive_sub_function: 3.002188205718994
		 	 - some_value: 150
		 	 - some_other_value: 1337

You can wrap any number of functions with the measure decorator, or use it as a context manager:

with measure("something_expensive"):
    SomeModel.objects.all().delete()

All measurements taken within a request/response cycle are fed to a signal called metrics_received. Currently the library only prints these metrics to the console, but in the future it will be expanded to log them to external services like InfluxDB.