12 Jun 09 Django and Python Logging
I started working on a simple Django application for accepting and recording Paypal donations. While I was working on the IPN code, it suddenly occurred to me that I really needed a way to log any errors that might occur. After all the IPN process will be initiated by Paypal completely out of my control (not counting the Paypal sandbox) and without any visual feedback. Thus, I’d like a record of the path through my code to make sure everything is working the way I expected.
I had learned about the brilliant Python logging module some time ago, and had even used it in my IRC bot application. But could you use this with Django?
I did some research and found a couple of blog posts and related Django projects. After studying them I came to the conclusion that indeed the Python logging module is an excellent way to add logging to your web application. And after working with it again, I am very impressed by the functionality that it offers and how easy it is to use. I would have killed to have something like this in my PHP days. So I laced some logging calls throughout my IPN listener code, and I’ll know soon enough if it works correctly. I’ll post more about this later. But for now, I’d like to add some links to the things I found useful related to logging in Django.
First there is this very useful blog post by Simon Willison titled “Debugging Django.” In addition to talking about logging, Simon also has tips on using the Python debugger, asserts, and some useful middleware.
Second, there is a Django application called django-logging. This application seems mainly aimed at getting your logging statements displayed at the bottom of your web pages while debugging a problem. Of course you could also hook in a logger to log to a file, which is more of what I wanted to do. Another useful feature of this application is that you can configure it to automatically log your application’s SQL queries.
And finally I looked at Fairview Computing’s Django request logging, or drlog. This project provides some middleware which adds the ability to add a unique identifier to each log entry to associate it with a given HTTP request. This allows you to easily trace a single request, even while multiple concurrent requests are happening.
Studying Simon’s blog post and the source code for the above two applications was very enlightening. In the end I decided to start with the bare Python logging facility for now, configuring it to write to a file. I was reassured to read that the Python logging module is thread safe. If I start using the logging module heavily, I may use the drlog middleware to help me map log statements to HTTP requests.
