APSL / django-profile-middleware

An easy to use customizable django profiling middleware to profile and find bottlenecks in your code and custom middlewares

django-profile-middleware

An easy to use customizable django profiling middleware for Django VERSION 1.10 or above. The code of this fork has been updated from the original repository in order to be able tu run with the newest python and django versions. This work couldn't be done without Himanshu Goyal's work.

Requirement

Django VERSION 1.10 or above and python 3.x

Installation

$ pip install git+https://github.com/csalom/django-profile-middleware.git#egg=django-profile-middleware

Middleware

Motivation: To provide an easy to use cProfile middleware to find bottlenecks in code.

How To Use (after installation)

In your settings add:

django_profile_middleware.middleware.ProfilerMiddleware to the end your MIDDLEWARE_CLASSES and django_profile_middleware to your INSTALLED_APPS

In case you also want to profile any of your custom middleware, just add all those middlewares below this one.

For example:

MIDDLEWARE = [
    'django.middleware.security.SecurityMiddleware',
    'django.contrib.sessions.middleware.SessionMiddleware',
    ....

    'django_profile_middleware.middleware.ProfilerMiddleware',

]

INSTALLED_APPS = [
    ....
    'django_profile_middleware',
    ....
] 

Enable

Add PROFILER in project's settings.py and set enable = True. Also make sure project running in DEBUG mode DEBUG = True

DEBUG = True

PROFILER = {
    'enable': True,
}

And

Profiling results will be printed on the console and will also be saved in a profiling_results.txt file. profiling_results.txt file will be created if not present and updated with the results at end if a file with same name already exists.

Example

        197 function calls (192 primitive calls) in 0.002 seconds

   Ordered by: internal time

   ncalls  tottime  percall  cumtime  percall filename:lineno(function)
   ...      ...      ...       ...     ...           ...
   ...      ...      ...       ...     ...           ...

Customize

You can customize the Profiler settings via adding optional fields to PROFILER key in settings.py

Default are

PROFILER = {
    'enable': True,

    #optional fields
    'sort': 'time',
    'count': None ,
    'output': ['console','file'],             
    'file_location': 'profiling_results.txt'

}

Description of optional fields

enable

Set this key to True to enable Profiler

'enable': True

To disable set to False

'enable': False

sort

Sort according to the set value. Default is 'time'. See documentaion for more options

count

Specify number of rows to output. By Default it will give all the rows.

output

Specify the form of output. Default is ['console',"file"]. In case only one of them is required just mention that in ou2tput field 'file' will write the file specified by 'file_location' field.

'output': ['console', 'file']  # default value

'output': ['console']

file_location

Specify the location of file you want to write in the results. Only valid if 'file' in 'output' field. Default value profiling_results.txt

Decorator

I have also added a decorator in the package which can be used optionally according to one's requirement.

Motivation: In case API takes very less time say 10 ms to execute, it is difficult to know which function is taking the maximum time, as, even the slowest one may take just 1 ms or even less . So this decorator provides you an option to run the whole code multiple times and thus scaling the total time to a more indicative value.

How To Use (after installation) In your views.py file add :

from django_profile_middleware.decorator import iterator

and at the function you want to run multiple time add decorator:

@iterator( *no. of times you want it to run )

for example:

@iterator(100)

Author

• Himanshu Goyal

Email me with any questions: [e-mail].

Updated by

• Carlos Salom

Email me with any questions: [e-mail].