Put open source to work
July 16–17, 2018: Training & Tutorials
July 18–19, 2018: Conference
Portland, OR

Herding cat pictures: How to develop, deploy, and operate services at Reddit scale

Greg Taylor (Reddit)
5:05pm5:45pm Thursday, July 19, 2018
Level: Beginner
Average rating: ****.
(4.71, 7 ratings)

Who is this presentation for?

  • Developers, managers, and curious individuals of all experience levels

Prerequisite knowledge

  • A basic understanding of the software development cycle and the team dynamics that support it

What you'll learn

  • Learn how a top-10 site overhauled its development and deploy cycle while rapidly growing
  • Gain a better understanding of some of the growing pains seen while scaling Reddit
  • Understand the need for developing, designing, and building with empathy for fellow developers

Description

For much of its existence, Reddit has been a relatively small engineering organization. Despite being one of the world’s busiest sites, there were many millions of users for every Reddit engineer.

The last few years have seen Reddit engineering more than triple in size (and ambition), leading to a significant change in process, tooling, and culture. Greg Taylor discusses Reddit’s evolution and explains how one of the world’s busiest sites develops, deploys, and operates services at significant scale. You’ll learn where things are today and get a preview of what the future holds for Reddit engineering.

Topics include:

  • The growing pains seen at a busy and rapidly expanding site
  • The shift from a single monolith to a service-oriented architecture
  • How Reddit develops and tests its dozens of services
  • Service ownership and accountability
  • The importance of developer empathy
  • How systems and services are deployed and operated at Reddit scale
Photo of Greg Taylor

Greg Taylor

Reddit

Greg Taylor leads Reddit’s Release Engineering Group within the company’s Infrastructure Division. He and his team primarily help Reddit engineers get their systems from concept to production. This involves developing tooling, building CI/CD pipelines, and a healthy serving of Kubernetes.