I just built a new linux 3.7.6 kernel and have a built-in module (which is not in use) throwing errors as fast as it probably can, ie, reporting a busy loop. Obviously this is not good and I have to deal with it, etc, but I would really -- really -- really -- really -- really -- really -- really appreciate it if rsyslog would not repeat the message on disk to the tune of 10MB/hour, etc.
I had thought syslog daemons prevented this by using "last message repeated" but that seems not to be the case. I have no idea what this option would even be called, since compression is a hot bed of references.
$RepeatedMessageReduction
and will try that later. It seems some years ago the rsyslog devs decided this feature (filtering duplicate messages) was not worth what it cost resource wise and may have depreciated it right out...