Logging performance improvment

[Avishay Balderman]

We have the log decorator that is used in diffrent modules: https://github.com/openstack/neutron/blob/master/neutron/common/log.py

We can add a check and see if the log is going to be written before we call it.

def log(method):
    """Decorator helping to log method calls."""
    def wrapper(*args, **kwargs):
        if LOG.isEnabledFor(logging.DEBUG):
            instance = args[0]
            data = {"class_name": (instance.__class__.__module__ + '.'
                                   + instance.__class__.__name__),
                    "method_name": method.__name__,
                    "args": args[1:], "kwargs": kwargs}
            LOG.debug(_('%(class_name)s method %(method_name)s'
                        ' called with arguments %(args)s %(kwargs)s'), data)
        return method(*args, **kwargs)
    return wrapper

See: https://docs.python.org/2/library/logging.html#logging.Logger.isEnabledFor

[Eugene Nikanorov (enikanorov)]

I have feeling that log library already capable of doing such check.

That's particularly is the reason to pass string substitution parameters as parameters to a logging functions instead of constructing log message in place.

[Launchpad Janitor (janitor)]

[Expired for neutron because there has been no activity for 60 days.]