I’m sure you’ve heard your co-workers say, “that’s an anti-pattern”. If you’re a systems thinker or work closely with large-scale distributed systems and open source technologies, you’ve most likely said those words yourself. And that is because anti-patterns exist everywhere, just like human error.

Spotting anti-patterns in any large scale production system is an artform, but it doesn’t need to be complex. Applying first principles of system design and common sense, in this talk I will describe a ruleset to guide you in identifying anti-patterns in your systems. While spotting deeply ingrained anti-patterns can require deep knowledge and expertise in the specific system, I believe that there is a way to generalize an approach that is good enough to get started. The first step to recovery is identification.

Taking Reddit’s infrastructure systems that serve well over hundreds of millions of users as an example, I will talk through anti-patterns that I work with everyday as an infrastructure engineer at the company. I will also talk through the long term impact that anti-patterns can have on your infrastructure systems, some of the cultural conflict that can exist around anti-patterns resolution, and how to address these challenges with grace.