Back to glossary
GLOSSARY · K

Kubernetes

The dominant open-source container orchestrator, the substrate most modern SRE work runs on.

Definition

Kubernetes (k8s) is an open-source container orchestrator originally designed at Google, now the default substrate for running containerized applications at scale. It schedules containers across a cluster of nodes, handles service discovery, autoscaling, rolling updates, secret management, persistent storage, and self-healing for failed pods. SRE teams care about Kubernetes because it both solves a category of operational pain (manual deploys, capacity, restarts) and creates a new one (the cluster itself becomes a complex distributed system to operate).

Why it matters

Most modern SRE jobs assume Kubernetes literacy because most modern infrastructure is on Kubernetes. A team's incident readiness is now measured in part by 'can the on-call engineer drain a node, scale a deployment, exec into a pod, and read events at 3am'. Investing in agent-driven runbooks for Kubernetes patterns is the single biggest MTTR cut available to most teams.

How Nova handles it

See the part of the platform that handles kubernetes in production.

Nova on Kubernetes