Authors: Brian Grant (Principal Engineer, Google) and Tim Hockin (Principal Engineer, Google)

We are humbled to be recognized by the community with this award.

We had high hopes when we created Kubernetes. We wanted to change the way cloud applications were deployed and managed. Whether we’d succeed or not was very uncertain. And look how far we’ve come in such a short time.

The core technology behind Kubernetes was informed by lessons learned from Google’s internal infrastructure, but nobody can deny the enormous role of the Kubernetes community in the success of the project. The community, of which Google is a part, now drives every aspect of the project: the design, development, testing, documentation, releases, and more. That is what makes Kubernetes fly.

While we actively sought partnerships and community engagement, none of us anticipated just how important the open-source community would be, how fast it would grow, or how large it would become. Honestly, we really didn’t have much of a plan.

We looked to other open-source projects for inspiration and advice: Docker (now Moby), other open-source projects at Google such as Angular and Go, the Apache Software Foundation, OpenStack, Node.js, Linux, and others. But it became clear that there was no clear-cut recipe we could follow. So we winged it.

Rather than rehashing history, we thought we’d share two high-level lessons we learned along the way.

First, in order to succeed, community health and growth needs to be treated as a top priority. It’s hard, and it is time-consuming. It requires attention to both internal project dynamics and outreach, as well as constant vigilance to build and sustain relationships, be inclusive, maintain open communication, and remain responsive to contributors and users. Growing existing contributors and onboarding new ones is critical to sustaining project growth, but that takes time and energy that might otherwise be spent on development. These things have to become core values in order for contributors to keep them going.

Second, start simple with how the project is organized and operated, but be ready to adopt to more scalable approaches as it grows. Over time, Kubernetes has transitioned from what was effectively a single team and git repository to many subgroups (Special Interest Groups and Working Groups), sub-projects, and repositories. From manual processes to fully automated ones. From informal policies to formal governance.

We certainly didn’t get everything right or always adapt quickly enough, and we constantly struggle with scale. At this point, Kubernetes has more than 20,000 contributors and is approaching one million comments on its issues and pull requests, making it one of the fastest moving projects in the history of open source.

Thank you to all our contributors and to all the users who’ve stuck with us on the sometimes bumpy journey. This project would not be what it is today without the community.