spectrum-python 0.9.8

Overview

Spectrum is a application that helps developers filter and sift through logs while developing or debugging locally. It accepts logging information via a REST API, syslog, or can be set to tail local files. All of these logging sources are then easily shown or hidden depending on the current needs of the developer.

This Python package makes it trivially easy to push logs from Python and/or Django into Spectrum.

Python

Installing Python support for Spectrum is as easy as:

$ pip install spectrum-python

Setting up a logging handler is also equally easy:

import logging
from spectrum.handlers import Spectrum

logger = logging.getLogger(__name__)
logger.setLevel(logging.DEBUG)

spectrum = Spectrum('my-logging-sublevel')
logger.addHandler(spectrum)

for i in range(5):
    logger.info("This would be sent as INFO.my-logging-sublevel")
    logger.warn("This would be sent as WARNING.my-logging-sublevel")
    logger.debug("This would be sent as DEBUG.my-logging-sublevel")

Django

To direct all logging from your Django project to Spectrum, you can use the predefined FIRE_HOSE logging configuration in your settings:

# settings.py

from spectrum.django import FIRE_HOSE
LOGGING = FIRE_HOSE

If you use celery, you’ll have one more setting to add:

CELERYD_HIJACK_ROOT_LOGGER = False

You can also use the fire_hose convenience method to quickly modify any of the pre-built logging configuration dict:

from spectrum.django import fire_hose


LOGGING = fire_hose()

LOGGING = fire_hose(log_db=False)

LOGGING = fire_hose(levels=(
    ('my.overly.verbose.module', 'WARNING'),
    ('some.other.module', 'CRITICAL'),
)

LOGGING = fire_hose(handler_kwargs={'url': '127.0.0.1:12345'})

If you prefer a more granular approach, you can configure specific handlers:

# settings.py

LOGGING = {
    ...
    'filters': {
        'request_id': {
            '()': 'spectrum.filters.RequestIdFilter'
        }
    },
    'handlers': {
        'myloggername': {
            'level': 'DEBUG',
            'class': 'spectrum.handlers.RestSpectrum',
            'sublevel': 'myloggername',
            'filters': ['request_id']
        }
    }
}

Make sure you include the request_id filter to filter out spectrum’s own requests from the other. Otherwise, Spectrum’s requests will be logged by Python, and those logs will in turn generate their own Spectrum request, causing and infinite recursion.