Skip to content

Latest commit

 

History

History
264 lines (201 loc) · 9.23 KB

README.md

File metadata and controls

264 lines (201 loc) · 9.23 KB
Raw

drf-api-tracking

build-status-image pypi-version Requirements Status Coverage Status

Overview

drf-api-tracking provides a Django model and DRF view mixin that work together to log Django Rest Framework requests to the database. You'll get these attributes for every request/response cycle to a view that uses the mixin:

Model field name Description Model field type
user User if authenticated, None if not Foreign Key
username_persistent Static field that persists the username even if the User model object is deleted CharField
requested_at Date-time that the request was made DateTimeField
response_ms Number of milliseconds spent in view code PositiveIntegerField
path Target URI of the request, e.g., "/api/" CharField
view Target VIEW of the request, e.g., "views.api.ApiView" CharField
view_method Target METHOD of the VIEW of the request, e.g., "get" CharField
remote_addr IP address where the request originated (X_FORWARDED_FOR if available, REMOTE_ADDR if not), e.g., "127.0.0.1" GenericIPAddressField
host Originating host of the request, e.g., "example.com" URLField
method HTTP method, e.g., "GET" CharField
query_params Dictionary of request query parameters, as text TextField
data Dictionary of POST data (JSON or form), as text TextField
response JSON response data TextField
status_code HTTP status code, e.g., 200 or 404 PositiveIntegerField

Requirements

  • Django 1.11, 2.0, 2.1, 2.2, 3.0
  • Django REST Framework and Python release supporting the version of Django you are using
Django Python DRF
1.11 2.7, 3.5, 3.6 3.5, 3.6, 3.7, 3.8, 3.9
2.0 3.5, 3.6, 3.7 3.7, 3.8, 3.9
2.1 3.5, 3.6, 3.7, 3.8 3.7, 3.8, 3.9
2.2 3.5, 3.6, 3.7, 3.8 3.7, 3.8, 3.9
3.0 3.5, 3.6, 3.7, 3.8 3.7, 3.8, 3.9

Installation

Install using pip...

$ pip install drf-api-tracking

Register with your Django project by adding rest_framework_tracking to the INSTALLED_APPS list in your project's settings.py file. Then run the migrations for the APIRequestLog model:

$ python manage.py migrate

Usage

Add the rest_framework_tracking.mixins.LoggingMixin to any DRF view to create an instance of APIRequestLog every time the view is called.

For instance:

# views.py
from rest_framework import generics
from rest_framework.response import Response
from rest_framework_tracking.mixins import LoggingMixin

class LoggingView(LoggingMixin, generics.GenericAPIView):
    def get(self, request):
        return Response('with logging')

For performance enhancement, explicitly choose methods to be logged using logging_methods attribute:

class LoggingView(LoggingMixin, generics.CreateModelMixin, generics.GenericAPIView):
    logging_methods = ['POST', 'PUT']
    model = ...

Moreover, you could define your own rules by overriding should_log method. If should_log evaluates to True a log is created.

class LoggingView(LoggingMixin, generics.GenericAPIView):
    def should_log(self, request, response):
        """Log only errors"""
        return response.status_code >= 400

At the example above, logging_methods attribute will be ignored. If you want to provide some extra rules on top of the http method filtering you should rewrite the should_log method.

class LoggingView(LoggingMixin, generics.GenericAPIView):
    def should_log(self, request, response):
        """Log only errors with respect on `logging_methods` attributes"""
        should_log_method = super(LoggingView, self).should_log(request, response)
        if not should_log_method:
            return False
        return response.status_code >= 400

A bit simpler.

class LoggingView(LoggingMixin, generics.GenericAPIView):
    def should_log(self, request, response):
        """Log only errors with respect on `logging_methods` attributes"""
        if not request.method in self.logging_methods:
            return False
        return response.status_code >= 400

Finally, you can also apply your customizations by overriding handle_log method. By default, all requests that satisfy should_log method are saved on the database.

class LoggingView(LoggingMixin, generics.GenericAPIView):
    def handle_log(self):
        # Do some stuff before saving.
        super(MockCustomLogHandlerView, self).handle_log()
        # Do some stuff after saving.

Though, you could define your own handling. For example save on an in-memory data structure store, remote logging system etc.

class LoggingView(LoggingMixin, generics.GenericAPIView):

    def handle_log(self):
        cache.set('my_key', self.log, 86400)

Or you could omit save a request to the database. For example,

class LoggingView(LoggingMixin, generics.GenericAPIView):
    def handle_log(self):
        """
        Save only very slow requests. Requests that took more than a second.
        """
        if self.log['response_ms'] > 1000:
            super(MockCustomLogHandlerView, self).handle_log()

If your endpoint accepts large file uploads, drf-api-tracking's default behavior to decode the request body may cause a RequestDataTooBig exception. This behavior can be disabled globally by setting DRF_TRACKING_DECODE_REQUEST_BODY = False in your settings.pyfile.

You can also customize this behavior for individual views by setting the decode_request_body attribute:

class LoggingView(LoggingMixin, generics.GenericAPIView):
    decode_request_body = False

Security

By default drf-api-tracking is hiding the values of those fields {'api', 'token', 'key', 'secret', 'password', 'signature'}. The default list hast been taken from Django itself (https://github.com/django/django/blob/stable/1.11.x/django/contrib/auth/init.py#L50).

You can complete this list with your own list by putting the fields you want to be hidden in the sensitive_fields parameter of your view.

class LoggingView(LoggingMixin, generics.CreateModelMixin, generics.GenericAPIView):
    sensitive_fields = {'my_secret_key', 'my_secret_recipe'}

By default drf-tracking allows API request log entries to be modified from Django admin. This can present a data integrity issue in production environments. In order to change this behavior, you can set DRF_TRACKING_ADMIN_LOG_READONLY to True in your settings.py file.

Development

In the folder there is a sample drf project: drf_api_sample if changes are made to this packages models, use this project to help generate new migrations, which should be checked in.

Testing

Install testing requirements.

$ pip install -r requirements.txt

Run with runtests.

$ ./runtests.py

You can also use the excellent tox testing tool to run the tests against all supported versions of Python and Django. Install tox globally, and then simply run:

$ tox

you can also use pyenv to install multiple versions of python and ensure they are found by tox by issuing:

pyenv install 3.8.4
pyenv install 3.7.7
pyenv install 3.6.11
pyenv local 3.8.4 3.7.7 3.6.11
pyenv global 3.8.4 3.7.7 3.6.11

Also ensure that you don't have a virtualenv activated when you run the tests else you might get the following error, or similar: ERROR: InterpreterNotFound: python3.6

Contributing

In order to make changes to the package itself, providing migrations or something else, make sure to install the current package with pip, otherwise using the drf_api_sample won't work.

pip install -e .

After this, you can edit models and creating migrations with

python drf_api_sample/manage.py makemigrations

Documentation

To build the documentation, you'll need to install mkdocs.

$ pip install mkdocs

To preview the documentation:

$ mkdocs serve
Running at: http://127.0.0.1:8000/

To build the documentation:

$ mkdocs build

travis

Install RVM to have a local user version of ruby/gem: https://rvm.io/rvm/install Then install travis like this: gem install travis add your secret key as per the link below: https://docs.travis-ci.com/user/encryption-keys/

pyenv

using pyenv you can install multiple versions of python so that tox can run tests against all installed versions of python

pyenv global 3.6.8 3.7.7 3.8.2

ensure that before running tox you don't have a virtualenv created and tox has been installed globally or via pipx