Pawel Zadrozny


About confusing or poorly documented features of tools and libraries I use.

Sentry with Django and Celery

19 Feb 2018 development | #python #django #sentry #queue #logging

Sentry is a service for errors log gathering and aggregation. Is available as paid SaaS or free to use standalone version. Read more on

Install Sentry SDK

Sentry comes with ready to use SDK package compatible with Python 2.7 and 3.x.

(venv) $ pip install raven; pip freeze | grep raven >> requirements.txt

or with pipenv:

$ pipenv install raven

Sentry minimal config for Django

In django add:

import raven

    'environment': 'production',  # optional but very useful
    'release': raven.fetch_git_sha(BASE_DIR),  # optional but very useful
    'dsn': 'https://example:[email protected]/n',  # DSN can be obtained from sentry panel

From now on Sentry will capture all unexpected exceptions, which are not catched in try-except block. Also all manually created log entries will be sent to Sentry. However this is limited only to exceptions raised by WSGI.

Catch exceptions raised inside Celery tasks

To make Sentry to be aware of Celery, register Sentry client just below Celery configuration. I assume you have created separate file for Celery config. If not i encourage you to read how to do it on: First steps with Celery 4 and Django

from raven.contrib.celery import register_logger_signal, register_signal
from raven.contrib.django.raven_compat.models import client

register_signal(client, ignore_expected=True)