Using Django 1.5.1:
DEBUG = False LOGGING = { 'version': 1, 'disable_existing_loggers': True, 'formatters': { 'verbose': { 'format': '%(levelname)s %(asctime)s %(module)s %(message)s' }, }, 'handlers': { 'console': { 'level': 'DEBUG', 'class': 'logging.StreamHandler', 'formatter': 'verbose', }, }, 'loggers': { # root logger '': { 'handlers': ['console'], }, #'django.request': { # 'handlers': ['console'], # 'level': 'DEBUG', # 'propagate': False, #}, } }
If I uncomment the commented lines and call a view which has 1/0
, the traceback is printed to the console:
ERROR 2013-11-29 13:33:23,102 base Internal Server Error: /comment/*******/ Traceback (most recent call last): ... File "*****/comments/views.py", line 10, in post 1/0 ZeroDivisionError: integer division or modulo by zero WARNING 2013-11-29 13:33:23,103 csrf Forbidden (CSRF cookie not set.): /comment/******/ [29/Nov/2013 13:33:23] "POST /comment/******/ HTTP/1.0" 500 27
But if the lines stay commented, no traceback is printed to the console, just:
[29/Nov/2013 13:33:23] "POST /comment/******/ HTTP/1.0" 500 27
I thought if django.request
logger is not configured, it would propagate to the root logger, which prints everything to console.
I didn't find any information that django.request
is special.
Why it doesn't work?
Here I read:
Prior to Django 1.5, the LOGGING setting always overwrote the default Django logging configuration. From Django 1.5 forward, it is possible to get the project’s logging configuration merged with Django’s defaults, hence you can decide if you want to add to, or replace the existing configuration.
If the disable_existing_loggers key in the LOGGING dictConfig is set to True (which is the default) the default configuration is completely overridden. Alternatively you can redefine some or all of the loggers by setting disable_existing_loggers to False.
In django/utils/log.py
:
# Default logging for Django. This sends an email to the site admins on every # HTTP 500 error. Depending on DEBUG, all other log records are either sent to # the console (DEBUG=True) or discarded by mean of the NullHandler (DEBUG=False). DEFAULT_LOGGING = { 'version': 1, 'disable_existing_loggers': False, 'filters': { 'require_debug_false': { '()': 'django.utils.log.RequireDebugFalse', }, 'require_debug_true': { '()': 'django.utils.log.RequireDebugTrue', }, }, 'handlers': { 'console':{ 'level': 'INFO', 'filters': ['require_debug_true'], 'class': 'logging.StreamHandler', }, 'null': { 'class': 'django.utils.log.NullHandler', }, 'mail_admins': { 'level': 'ERROR', 'filters': ['require_debug_false'], 'class': 'django.utils.log.AdminEmailHandler' } }, 'loggers': { 'django': { 'handlers': ['console'], }, 'django.request': { 'handlers': ['mail_admins'], 'level': 'ERROR', 'propagate': False, }, 'py.warnings': { 'handlers': ['console'], }, } }
So by default django.request
has propagate = False
. But In my case I have 'disable_existing_loggers': True
.
The solution is to prevent Django from configuring logging and handle it ourselves. Fortunately this is easy. In settings.py
:
LOGGING_CONFIG = None LOGGING = {...} # whatever you want, as you already have import logging.config logging.config.dictConfig(LOGGING)
UPDATE ~March 2015: Django has clarified their documentation:
If the disable_existing_loggers key in the LOGGING dictConfig is set to True then all loggers from the default configuration will be disabled. Disabled loggers are not the same as removed; the logger will still exist, but will silently discard anything logged to it, not even propagating entries to a parent logger. Thus you should be very careful using 'disable_existing_loggers': True; it’s probably not what you want. Instead, you can set disable_existing_loggers to False and redefine some or all of the default loggers; or you can set LOGGING_CONFIG to None and handle logging config yourself.
For posterity and detail: The explanation? Most of the confusion I think comes down to Django's poor explanation of disable_existing_loggers
, which says that when True, "the default configuration is completely overridden". In your own answer you discovered that is not correct; what's happening is that the existing loggers, which Django already configures, are disabled not replaced.
The Python logging documentation explains it better (emphasis added):
disable_existing_loggers – If specified as False, loggers which exist when this call is made are left alone. The default is True because this enables old behaviour in a backward-compatible way. This behaviour is to disable any existing loggers unless they or their ancestors are explicitly named in the logging configuration.
Based on Django docs we think, "override the defaults with my own LOGGING configuration and anything I don't specify will bubble up". I've tripped over this expectation as well. The behavior we expect is along the lines of replace_existing_loggers (which isn't a real thing). Instead the Django loggers are shut up not bubbled up.
We need to prevent the setup of these Django loggers in the first place and here the Django docs are more helpful:
If you don’t want to configure logging at all (or you want to manually configure logging using your own approach), you can set LOGGING_CONFIG to None. This will disable the configuration process.
Note: Setting LOGGING_CONFIG to None only means that the configuration process is disabled, not logging itself. If you disable the configuration process, Django will still make logging calls, falling back to whatever default logging behavior is defined.
Django will still use its loggers but since they are not handled (and then disabled) by the configuration, those loggers will bubble up as expected. A simple test with the above settings:
manage.py shell >>> import logging >>> logging.warning('root logger') WARNING 2014-03-11 13:35:08,832 root root logger >>> l = logging.getLogger('django.request') >>> l.warning('request logger') WARNING 2014-03-11 13:38:22,000 django.request request logger >>> l.propagate, l.disabled (1, 0)
Ok, so the behavior is "correct", but not expected. django/conf/__init__.py:65
:
def _configure_logging(self): ... if self.LOGGING_CONFIG: from django.utils.log import DEFAULT_LOGGING # First find the logging configuration function ... logging_config_path, logging_config_func_name = self.LOGGING_CONFIG.rsplit('.', 1) logging_config_module = importlib.import_module(logging_config_path) logging_config_func = getattr(logging_config_module, logging_config_func_name) logging_config_func(DEFAULT_LOGGING) if self.LOGGING: # Backwards-compatibility shim for #16288 fix compat_patch_logging_config(self.LOGGING) # ... then invoke it with the logging settings logging_config_func(self.LOGGING)
What is happening is that default logging configuration is applied and django.request
logger is created. Then my custom LOGGING
configuration is applied with disable_existing_loggers = True
, but Python doesn't delete already existing logger django.request
, but only disables it.
So I have to manually reconfigure django.request
logger in my configuration. :(
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With