Site Reliability Engineering

Chapter 31. Communication and Collaboration in SRE

Written by Niall Murphy with Alex Rodriguez, Carl Crous, Dario Freni, Dylan Curley, Lorenzo Blanco, and Todd Underwood

Edited by Betsy Beyer

The organizational position of SRE in Google is interesting, and has effects on how we communicate and collaborate.

To begin with, there is a tremendous diversity in what SRE does, and how we do it. We have infrastructural teams, service teams, and horizontal product teams. We have relationships with product development teams ranging from teams that are many times our size, to teams roughly the same size as their counterparts, and situations in which we are the product development team. SRE teams are made up of people with systems engineering or architectural skills (see [Hix15b]), software engineering skills, project management skills, leadership instincts, backgrounds in all kinds of industries (see Chapter 33), and so on. We don’t have just one model, and we have found a variety of configurations that work; this flexibility suits our ultimately pragmatic nature.

It’s also true that SRE is not a command-and-control organization. Generally, we owe allegiance to at least two masters: for service or infrastructure SRE teams, we work closely with the corresponding product development teams that work on those services or that infrastructure; we also obviously work in the context of SRE generally. The service relationship is very strong, since we are held accountable for the performance of those ...