Suppress multiple messages with same content in Python logging module AKA log compression

By design, my application sometimes produces repeating errors which fill up the log file and make it annoying to read. It looks like that:

WARNING:__main__:CRON10: clock unset or no wind update received in 60 sec -> supressed rrd update
WARNING:__main__:CRON10: clock unset or no wind update received in 60 sec -> supressed rrd update
WARNING:__main__:CRON10: clock unset or no wind update received in 60 sec -> supressed rrd update
WARNING:__main__:CRON10: clock unset or no wind update received in 60 sec -> supressed rrd update

How can I use the Python logging module to suppress repeating messages and output something more rsyslog style (http://www.rsyslog.com/doc/rsconf1_repeatedmsgreduction.html):

WARNING:__main__:CRON10: clock unset or no wind update received in 60 sec -> supressed rrd update
--- The last message repeated 3 times

Is there a way to extend logging or do I have to write a completly own logger?

The code I use for logging is:

logging.basicConfig(format='%(asctime)s %(message)s')
logging.basicConfig(level=logging.info)
logger = logging.getLogger(__name__)
hdlr = logging.FileHandler(LOGFILE)
hdlr.setFormatter(formatter)
logger.addHandler(hdlr)

Any ideas on that?

Solution

You can create a logging.Filter that will keep track of the last logged record and filter out any repeated (similar) records, something like:

import logging

class DuplicateFilter(logging.Filter):

    def filter(self, record):
        # add other fields if you need more granular comparison, depends on your app
        current_log = (record.module, record.levelno, record.msg)
        if current_log != getattr(self, "last_log", None):
            self.last_log = current_log
            return True
        return False

Then just add it to the logger/handler you use (i.e. hdlr.addFilter(DuplicateFilter())) or the root logger to filter all default logs. Here's a simple test:

import logging

logging.warn("my test")
logging.warn("my repeated test")
logging.warn("my repeated test")
logging.warn("my repeated test")
logging.warn("my other test")

logger = logging.getLogger()  # get the root logger
logger.addFilter(DuplicateFilter())  # add the filter to it

logging.warn("my test")
logging.warn("my repeated test")
logging.warn("my repeated test")
logging.warn("my repeated test")
logging.warn("my other test")

This will print out:

WARNING:root:my test
WARNING:root:my repeated test
WARNING:root:my repeated test
WARNING:root:my repeated test
WARNING:root:my other test
WARNING:root:my test
WARNING:root:my repeated test
WARNING:root:my other test