Skip to content

Instantly share code, notes, and snippets.

Created September 22, 2010 13:47
Show Gist options
  • Save vsajip/591699 to your computer and use it in GitHub Desktop.
Save vsajip/591699 to your computer and use it in GitHub Desktop.
import logging
import Queue as queue
except ImportError:
import queue
import threading
class QueueHandler(logging.Handler):
This handler sends events to a queue. Typically, it would be used together
with a multiprocessing Queue to centralise logging to file in one process
(in a multi-process application), so as to avoid file write contention
between processes.
This code is new in Python 3.2, but this class can be copy pasted into
user code for use with earlier Python versions.
def __init__(self, queue):
Initialise an instance, using the passed queue.
self.queue = queue
def enqueue(self, record):
Enqueue a record.
The base implementation uses put_nowait. You may want to override
this method if you want to use blocking, timeouts or custom queue
def prepare(self, record):
Prepares a record for queuing. The object returned by this method is
The base implementation formats the record to merge the message
and arguments, and removes unpickleable items from the record
You might want to override this method if you want to convert
the record to a dict or JSON string, or send a modified copy
of the record while leaving the original intact.
# The format operation gets traceback text into record.exc_text
# (if there's exception data), and also puts the message into
# record.message. We can then use this to replace the original
# msg + args, as these might be unpickleable. We also zap the
# exc_info attribute, as it's no longer needed and, if not None,
# will typically not be pickleable.
record.msg = record.message
record.args = None
record.exc_info = None
return record
def emit(self, record):
Emit a record.
Writes the LogRecord to the queue, preparing it for pickling first.
except (KeyboardInterrupt, SystemExit):
class QueueListener(object):
This class implements an internal threaded listener which watches for
LogRecords being added to a queue, removes them and passes them to a
list of handlers for processing.
_sentinel = None
def __init__(self, queue, *handlers):
Initialise an instance with the specified queue and
self.queue = queue
self.handlers = handlers
self._stop = threading.Event()
self._thread = None
def dequeue(self, block):
Dequeue a record and return it, optionally blocking.
The base implementation uses get. You may want to override this method
if you want to use timeouts or work with custom queue implementations.
return self.queue.get(block)
def start(self):
Start the listener.
This starts up a background thread to monitor the queue for
LogRecords to process.
self._thread = t = threading.Thread(target=self._monitor)
def prepare(self , record):
Prepare a record for handling.
This method just returns the passed-in record. You may want to
override this method if you need to do any custom marshalling or
manipulation of the record before passing it to the handlers.
return record
def handle(self, record):
Handle a record.
This just loops through the handlers offering them the record
to handle.
record = self.prepare(record)
for handler in self.handlers:
def _monitor(self):
Monitor the queue for records, and ask the handler
to deal with them.
This method runs on a separate, internal thread.
The thread will terminate if it sees a sentinel object in the queue.
q = self.queue
has_task_done = hasattr(q, 'task_done')
while not self._stop.isSet():
record = self.dequeue(True)
if record is self._sentinel:
if has_task_done:
except queue.Empty:
# There might still be records in the queue.
while True:
record = self.dequeue(False)
if record is self._sentinel:
if has_task_done:
except queue.Empty:
def stop(self):
Stop the listener.
This asks the thread to terminate, and then waits for it to do so.
Note that if you don't call this before your application exits, there
may be some records still left on the queue, which won't be processed.
self._thread = None
def main():
q = queue.Queue(-1)
qh = QueueHandler(q)
h = logging.StreamHandler()
ql = QueueListener(q, h)
root = logging.getLogger()
f = logging.Formatter('%(threadName)s: %(message)s')
# The log output will display the thread which generated
# the event (the main thread) rather than the internal
# thread which monitors the internal queue. This is what
# you want to happen.
root.warning('Look out!')
if __name__ == '__main__':
Copy link

sidazad commented Feb 19, 2016

Hi - thanks for posting this. I experimented with this and found that it takes an enormous amount of time (18 seconds or so for 100,000 prints to a FileHandler) as opposed to 7-8 seconds a single threaded logger. I understand there are costs of synchronization but this seems a bit much.
Any thoughts?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment