zlacker

Yeah to me it doesn't make any sense - configuration changes are just as likely to break stuff (as they've discovered the hard way) and both of these issues could have been found in a testing environment before being deployed to production