Building a Kubernetes Platform for Trillion Row Tables– CN Data at Scale

Kubernetes is an amazing platform for data, but it requires focus to deliver production apps, especially if they are large. In this talk, I’ll share learnings from operating a database SaaS for ClickHouse as well as helping customers do it themselves. We’ll describe some of the attributes of large analytic applications, and then discuss common approaches to building platforms to operate them. Our talk will cover the use of standard tools like Terraform, Helm, and Argo CD, which work together to bring up Kubernetes and the apps that run inside it. Along the way, we’ll describe various “aha!” experiences, tools, and even team organization to support large database applications on Kubernetes.