The Production Engineering Lifecycle: How We Build, Run, and Disband Great Reliability-focused Teams

Engineers focused on reliability and scalability under real-world conditions are a scarce resource in any organization. How do we know where to deploy them, and how do we use them in the best possible way? In Facebook's Production Engineering team, we have this problem all the time, and we've dealt with it a variety of ways throughout the years. Some of these ways have worked better than others, and we'd like to share what works and what hasn't.

In this talk, we will share our approaches to when to start a production engineering team, how to integrate that team into the existing development team, how to prioritize and divide work between engineers, and even when to disband or merge the team. We will also discuss practical matters such as how we divide on call responsibilities and roadmap items, and how we integrate engineers in multiple locations and time zones. 

Andrew has been a Production Engineer at Facebook since 2009, as a tech lead for a number of teams, supporting systems such as Hadoop and Memcache. Before that, he spent eight years managing SaaS deployments at CollabNet, and various sysadmin positions before that.