I am trying to debug a python script, but don't see any data in python.log. What can I do to get more information from python.log?

[jbsplunk]

I am attempting to do some debugging on a python script I've written to execute as an alert script. I've tried to edit the python logging settings in log.cfg but am unable to get any data that is useful in debugging the problem. Can I do something to alter the behavior so that I can get more information? Right now, all I see messages like this:

2014-02-04 11:22:19,552 -0400 INFO runshellscript:129 - 

[jrodman · Accepted solution]

Generally, it probably makes sense to review what the other scripts that ship with splunk are doing. For example, sendemail.py is the most frequently used alert in the system. It arranges to log messages to python.log in the following way:

import (...) splunk.mining.dcutils as dcu
[...]
logger = dcu.getLogger()
[...]
logger.error(error_message_string)

Personally I'd do a slightly simpler

import splunk.mining.dcutils
logger = splunk.mining.dcutils.getLogger()

The way this is used within the script is to catch exceptions and then render them into log messages. eg

except Exception, e:
      logger.error("Could not get email credentials from splunk, using no credentials. Error: %s" % (str(e)))

If you need more context, such as the full exception, you can use normal python methods to render the exception to a string, eg

logger.error(traceback.format_exc())

see https://docs.python.org/2.7/library/traceback.html for more.

[awurster]

absolutely brilliant. this sort of topic has been answered so many different times.. but this response is just so elegant, accurate and fresh.