But how do we know what good automation is? It's easy to see terrible ops work, but it's hard to see amazing ops work. This leads many organizations to end up t

But how do we know what good automation is? It's easy to see terrible ops work, but it's hard to see amazing ops work. This leads many organizations to end up thinking that ops skill is about outage response. Guess what: The person that is best at fixing at outage is typically the one that caused it! Imagine I have two people in an ops team: One writes a lot of new code, and ships a lot of buggy features, and quickly reacts to the problems. People get the impression that they are working on something hard, and they are grateful for all the hard work. Another one takes longer to ship, but the automation just works. Therefore, whatever he was doing was easy, and they are just slow at doing it. So sadly, in most environments it pays off to actually have problems in your ops automation, for the visibility that you get by saving the day.