Documentation

Kubestead Documentation

Everything you need to install, configure, and operate Kubestead in your Kubernetes cluster.

Getting Started

Kubestead is a Kubernetes-native deployment controller that automates canary rollouts, metric analysis, and rollback decisions. Follow the links below to get started.

Quickstart Install the controller and deploy your first canary rollout in 10 minutes. Rollout Spec Complete reference for all Rollout CRD fields, with YAML examples. Canary Analysis Define PromQL queries, thresholds, and evaluation windows. API Reference REST API for programmatic rollout management and integration.

Architecture Overview

Kubestead runs a single controller pod in your cluster. It watches for Rollout CRD resources, executes the canary steps you define, queries your metrics backend at each analysis interval, and issues promote or rollback commands automatically.

The controller requires no external control plane. All state is stored in Kubernetes etcd via CRD status fields. If the controller pod restarts, it reconciles from the current cluster state — no data loss, no mid-rollout disruption.

Prerequisites

  • Kubernetes 1.24+ (EKS, GKE, AKS, or self-managed)
  • Prometheus (or Datadog on Team/Scale tier) for metric analysis
  • Helm 3.x for installation