Logging uncaught exceptions in Python

How do you cause uncaught exceptions to output via the logging module rather than to stderr?

I realize the best way to do this would be:

try:
    raise Exception, 'Throwing a boring exception'
except Exception, e:
    logging.exception(e)

But my situation is such that it would be really nice if logging.exception(...) were invoked automatically whenever an exception isn't caught.

Here's a complete small example that also includes a few other tricks:

import sys
import logging
logger = logging.getLogger(__name__)
handler = logging.StreamHandler(stream=sys.stdout)
logger.addHandler(handler)

def handle_exception(exc_type, exc_value, exc_traceback):
    if issubclass(exc_type, KeyboardInterrupt):
        sys.__excepthook__(exc_type, exc_value, exc_traceback)
        return

    logger.error("Uncaught exception", exc_info=(exc_type, exc_value, exc_traceback))

sys.excepthook = handle_exception

if __name__ == "__main__":
    raise RuntimeError("Test unhandled")

• Ignore KeyboardInterrupt so a console python program can exit with Ctrl + C.

• Rely entirely on python's logging module for formatting the exception.

• Use a custom logger with an example handler. This one changes the unhandled exception to go to stdout rather than stderr, but you could add all sorts of handlers in this same style to the logger object.

As Ned pointed out, sys.excepthook is invoked every time an exception is raised and uncaught. The practical implication of this is that in your code you can override the default behavior of sys.excepthook to do whatever you want (including using logging.exception).

As a straw man example:

import sys
def foo(exctype, value, tb):
    print('My Error Information')
    print('Type:', exctype)
    print('Value:', value)
    print('Traceback:', tb)

Override sys.excepthook:

>>> sys.excepthook = foo

Commit obvious syntax error (leave out the colon) and get back custom error information:

>>> def bar(a, b)
My Error Information
Type: <type 'exceptions.SyntaxError'>
Value: invalid syntax (<stdin>, line 1)
Traceback: None

For more information about sys.excepthook, read the docs.

Why not:

import sys
import logging
import traceback

def log_except_hook(*exc_info):
    text = "".join(traceback.format_exception(*exc_info()))
    logging.error("Unhandled exception: %s", text)

sys.excepthook = log_except_hook

None()

Here is the output with sys.excepthook as seen above:

$ python tb.py
ERROR:root:Unhandled exception: Traceback (most recent call last):
  File "tb.py", line 11, in <module>
    None()
TypeError: 'NoneType' object is not callable

Here is the output with the sys.excepthook commented out:

$ python tb.py
Traceback (most recent call last):
  File "tb.py", line 11, in <module>
    None()
TypeError: 'NoneType' object is not callable

The only difference is that the former has ERROR:root:Unhandled exception: at the beginning of the first line.

The method sys.excepthook will be invoked if an exception is uncaught: http://docs.python.org/library/sys.html#sys.excepthook

> When an exception is raised and uncaught, the interpreter calls sys.excepthook with three arguments, the exception class, exception instance, and a traceback object. In an interactive session this happens just before control is returned to the prompt; in a Python program this happens just before the program exits. The handling of such top-level exceptions can be customized by assigning another three-argument function to sys.excepthook.

To build on Jacinda's answer, but using a logger object:

def catchException(logger, typ, value, traceback):
    logger.critical("My Error Information")
    logger.critical("Type: %s" % typ)
    logger.critical("Value: %s" % value)
    logger.critical("Traceback: %s" % traceback)

# Use a partially applied function
func = lambda typ, value, traceback: catchException(logger, typ, value, traceback)
sys.excepthook = func

In my case (using python 3) when using @Jacinda 's answer the content of the traceback was not printed. Instead, it just prints the object itself: <traceback object at 0x7f90299b7b90>.

Instead I do:

import sys
import logging
import traceback

def custom_excepthook(exc_type, exc_value, exc_traceback):
    # Do not print exception when user cancels the program
    if issubclass(exc_type, KeyboardInterrupt):
        sys.__excepthook__(exc_type, exc_value, exc_traceback)
        return

    logging.error("An uncaught exception occurred:")
    logging.error("Type: %s", exc_type)
    logging.error("Value: %s", exc_value)

    if exc_traceback:
        format_exception = traceback.format_tb(exc_traceback)
        for line in format_exception:
            logging.error(repr(line))

sys.excepthook = custom_excepthook

Wrap your app entry call in a try...except block so you'll be able to catch and log (and perhaps re-raise) all uncaught exceptions. E.g. instead of:

if __name__ == '__main__':
    main()

Do this:

if __name__ == '__main__':
    try:
        main()
    except Exception as e:
        logger.exception(e)
        raise

Although @gnu_lorien's answer gave me good starting point, my program crashes on first exception.

I came with a customised (and/or) improved solution, which silently logs Exceptions of functions that are decorated with @handle_error.

import logging

__author__ = 'ahmed'
logging.basicConfig(filename='error.log', level=logging.DEBUG)


def handle_exception(exc_type, exc_value, exc_traceback):
    import sys
    if issubclass(exc_type, KeyboardInterrupt):
        sys.__excepthook__(exc_type, exc_value, exc_traceback)
        return
    logging.critical(exc_value.message, exc_info=(exc_type, exc_value, exc_traceback))


def handle_error(func):
    import sys

    def __inner(*args, **kwargs):
        try:
            return func(*args, **kwargs)
        except Exception, e:
            exc_type, exc_value, exc_tb = sys.exc_info()
            handle_exception(exc_type, exc_value, exc_tb)
        finally:
            print(e.message)
    return __inner


@handle_error
def main():
    raise RuntimeError("RuntimeError")


if __name__ == "__main__":
    for _ in xrange(1, 20):
        main()

Maybe you could do something at the top of a module that redirects stderr to a file, and then logg that file at the bottom

sock = open('error.log', 'w')               
sys.stderr = sock

doSomething() #makes errors and they will log to error.log

logging.exception(open('error.log', 'r').read() )

To answer the question from Mr.Zeus discussed in the comment section of the accepted answer, I use this to log uncaught exceptions in an interactive console (tested with PyCharm 2018-2019). I found out sys.excepthook does not work in a python shell so I looked deeper and found that I could use sys.exc_info instead. However, sys.exc_info takes no arguments unlike sys.excepthook that takes 3 arguments.

Here, I use both sys.excepthook and sys.exc_info to log both exceptions in an interactive console and a script with a wrapper function. To attach a hook function to both functions, I have two different interfaces depending if arguments are given or not.

Here's the code:

def log_exception(exctype, value, traceback):
    logger.error("Uncaught exception occurred!",
                 exc_info=(exctype, value, traceback))


def attach_hook(hook_func, run_func):
    def inner(*args, **kwargs):
        if not (args or kwargs):
            # This condition is for sys.exc_info
            local_args = run_func()
            hook_func(*local_args)
        else:
            # This condition is for sys.excepthook
            hook_func(*args, **kwargs)
        return run_func(*args, **kwargs)
    return inner


sys.exc_info = attach_hook(log_exception, sys.exc_info)
sys.excepthook = attach_hook(log_exception, sys.excepthook)

The logging setup can be found in gnu_lorien's answer.