Sentry doesn't group logging properly

[memee]

Sentry should group messages logged in this code. The only variable is bundle.obj.order_number.
Instead it creates different groups based on bundle.obj.order_number

from __future__ import unicode_literals

[...]

try:
    fault_service = msg.FaultService.objects.get(
        header__order_number=bundle.obj.fault_number
    )
except msg.FaultService.DoesNotExist:
    logger.warning(
        "Couldn't find FaultService for FaultResource with "
        "order_number = %s" % bundle.obj.order_number
    )
    return None
else:
    return fault_service.pk

[dcramer]

Is this getsentry.com, if not, what version of Sentry?

On Thursday, February 7, 2013 at 10:37 PM, memee wrote:

Sentry should group messages logged in this code. The only variable is bundle.obj.order_number.
Instead it creates different groups based on bundle.obj.order_number
from future import unicode_literals [...] try: fault_service = msg.FaultService.objects.get( header__order_number=bundle.obj.fault_number ) except msg.FaultService.DoesNotExist: logger.warning( "Couldn't find FaultService for FaultResource with " "order_number = %s" % bundle.obj.order_number ) return None else: return fault_service.pk (http://fault_service.pk)

—
Reply to this email directly or view it on GitHub (#773).

[memee]

Sentry 5.3.3

[dcramer]

Actually reading this more, the default behavior for a non-exception is to group by message.

In this case, you'd want to change your call from:

logging.warning('foo bar %s' % 'baz')

To be the logging format capable:

logging.warning('foo bar %s', 'baz')

Sentry will handle the grouping of this like you're expecting

On Friday, February 8, 2013 at 1:31 AM, memee wrote:

Sentry 5.3.3

—
Reply to this email directly or view it on GitHub (#773 (comment)).

[memee]

Thank you, it worked.

[marjanmo]

Nice solution, thanks! However, is there a way to also aggregate/group messages, created with .format() method opposed to %s?

[dcramer]

@marjanmo not automatically as that is not supported by the python logging module. You should simply use the supported API if you want that behavior.