Running out of disk space (or inodes!) is probably how everyone running a server learns about monitoring and alerts