Django integrations docs should describe Django 1.3's LOGGING setting

[rassie]

In Django 1.3, it was a lot easier for me to just configure a SentryHandler in settings.py's LOGGING:

LOGGING = {
    .....
    'handlers': {
        ....
        'sentry': {
            'level': 'DEBUG',
            'class': 'sentry.client.handlers.SentryHandler',
            'formatter': 'verbose'
        },
        ...
   },

Then, configure log levels and just use the lines:

import logging
logger = logging.getLogger(__name__)

wherever logging is needed.

This ticket is a suggestion that maybe this kind of configuration should be documented more explicitely (unless some bigger problems exist which I'm not aware of).

[gtaylor]

I did the same with mine, works great. This would definitely be a good add.

[dcramer]

That looks great. I'll try to update the docs (unless someone beats me to it) as soon as I get a chance.

[kmike]

There is a gotcha with django 1.3 logging & sentry: sentry attaches sentry_exception_handler to django's got_request_exception signal (see https://github.com/dcramer/django-sentry/blob/master/sentry/client/models.py#L59). So the exception in request is sent to sentry anyway, even if SentryHandler is not in LOGGING. And if SentryHandler is added to LOGGING as a handler for 'django.request' logger or its parent (e.g. for root logger) then all request exceptions will be stored twice.

First solution is to set propagate=False for django.request logger (but that's not always what we want). Second solution is to remove the signal handler and rely on logging. The third option is to use a logging filter.

[mbaechtold]

rassie, maybe you can help me here.

I can't get Sentry to log my manual log-calls in Django 1.3. Logging of exceptions is working though.
Could you please show me the corresponding loggers to your handlers?

I tried to set up settings.HANDLERS and then in my view:

import logging
logger = logging.getLogger(__name__)
logger.debug('Some debug information',)

Do I still need to add the following (from the docs) to my settings.py? If I do so, my internal web server suddenly stops when I call the view that contains the log call.

import logging
from sentry.client.handlers import SentryHandler

logger = logging.getLogger()
# ensure we havent already registered the handler
if SentryHandler not in map(lambda x: x.__class__, logger.handlers):
    logger.addHandler(SentryHandler())

    # Add StreamHandler to sentry's default so you can catch missed exceptions
    logger = logging.getLogger('sentry.errors')
    logger.propagate = False
    logger.addHandler(logging.StreamHandler())

[rassie]

@Tini: the configuration example from the docs is unneeded, since setting LOGGERS in settings would do exactly the same. I started with this example, removed filters and anything that used them, added the handlers quoted above and added sentry to the default logger. After that, it just worked.

[mbaechtold]

@rassie: thank you. I'll try that. I hope this will soon find its way into the official docs.