Switch branches/tags
Nothing to show
Find file
Fetching contributors…
Cannot retrieve contributors at this time
71 lines (49 sloc) 2.41 KB


flask-tracking is a tracking app for Flask that logs HTTP request and response information in a capped MongoDB collection. For each request, the execution time is stored, making it easy to identify slow requests. flask-tracking supports Flask-Login to store the currently logged in user and requires MongoEngine to store the tracking information.


In your main application file, install the tracking app as follows:

from flask.ext.tracking import Tracking

Make sure to set up the tracking app as early as possible to ensure accurate logging of the response and execution time. If you're using Flask-gzip, install the tracking middleware afterwards to avoid logging compressed responses.

From now on, all requests are stored in the tracking collection.


  • TRACKING_EXCLUDE: A list of regular expressions of paths that you want to exclude from being tracked.


  • TRACKING_EXCLUDE_BODY: A list of regular expressions of paths where you don't want to track the request and response body. This is useful if specific paths receive or return sensitive data.


  • TRACKING_MAX_BODY_LENGTH: The maximum size in bytes of the request and response body that should be stored.


The following query shows all requests that took longer than one second to execute.

from flask.ext.tracking.documents import Tracking

The following query shows all requests that were served between 23:40 and 23:45 UTC on October 22th, 2012:

from flask.ext.tracking.documents import Tracking
Tracking.objects.filter(date_created__gte='2012-10-22 23:40:00', date_created__lte='2012-10-22 23:45:00') 

Storing custom data

Sometimes it is useful to store custom data in the tracking table for a given request. To do that, simply assign your data to request._tracking_data. Example:

from flask import request
def my_view():
    request._tracking_data = {'action: 'view', 'my_stuff': [1, 2, 3]}
    return render()

Custom data is stored in a DynamicField called custom_data and can therefore contain any information.