We love investing in small and scrappy teams, but I have noticed a distinct anti-pattern that has affected many of these teams: as their site or service grows they neglect ops and it comes back to bite them. I believe the source of this pattern is twofold. First, many folks today are perfectly capable doing a fair bit of ops themselves because they have had to do it and also because there is just a ton of information available on the web. I find myself doing that when I help Susan with DailyLit. Need to back up the database? Google around, pick a solution, make it work. Done. Or so it seems. The second source of this anti-pattern is a misunderstanding about what a good ops person is like. Too many folks seem to equate this with some very limited notion of a sys admin who can get a box installed and not much else. They conclude based on this perception that they don’t have enough work for someone like that and hence don’t need them. Instead they keep going with googling around and sort of making stuff work. Then one day something major happens and all hell breaks loose. My recommendation is that as soon as you get to a handful of machines as part of your production environment, hire someone who loves ops. This is very different from saying hire a sys admin. You want to find an engineer who absolutely can contribute to your development and help advance features, but if given half a chance will work on installing better server monitoring, improving the deployment process, looking for security holes, moving to point-in-time recovery for your database and so on. And yes, there are people who absolutely love this. Having a great ops person will make a huge difference. Because for anything that one might google, there are dozens of suggestions out there many of which are misleading if not downright wrong. A great ops person by contrast will either know the answer or spend the time to really dig in and get it right. The net result will be better service quality, faster development and less cost. The latter tends to be the case because a great ops person will take pride in getting the most out of whatever equipment you have. So this is one of those cases where with the right hiring you can have better, faster, cheaper.
