---
title: Kubernetes
description: Deploying Octopus to Kubernetes — a short overview that points to the full Kubernetes section for install, the in-process operator, Helm, probes, and EndpointSlice discovery.
---

# Kubernetes

Octopus is **Kubernetes-first**. The same binary you run with Docker can also run an
**in-process operator** that watches the cluster and programs its own router — no sidecar and no
separate control plane. It exposes dedicated health probes, drains gracefully on rolling
updates, and discovers upstream Pods through the EndpointSlice API.

<Callout type="info">
  This page is a pointer. The full, maintained reference lives in the
  **[Kubernetes section](/docs/octopus/kubernetes)** — install, operator, Helm, probes, discovery, and
  cert-manager TLS. Start there.
</Callout>

## Two ways to install

Both are documented under [Kubernetes → Install](/docs/octopus/kubernetes/install).

- **Helm (recommended)** — a parameterized, upgradable install with RBAC, probes, autoscaling,
  a `PodDisruptionBudget`, and an optional `ServiceMonitor`:

  ```sh
  helm install octopus deploy/helm/octopus \
    --namespace octopus --create-namespace
  ```

- **Raw manifests** — a single `kubectl apply` for a Helm-free install, using the bundled
  `deploy/kubernetes/octopus.yaml`.

Routes, upstreams, plugins, and auth go under the chart's `config` value, which is rendered into
a ConfigMap and mounted at `/etc/octopus/config.yaml` — the same file the
[container image](/docs/octopus/deployment/docker) expects.

## What the Kubernetes section covers

<Cards>
  <Card title="Install" description="Helm and raw manifests, CRD installation, and Gateway API prerequisites." href="/docs/octopus/kubernetes/install" />
  <Card title="Operator" description="The in-process controller: watchers, reconcile pipeline, precedence, and leader election." href="/docs/octopus/kubernetes/operator" />
  <Card title="Helm chart" description="The values reference — replicas, RBAC, HPA, PDB, and ServiceMonitor." href="/docs/octopus/kubernetes/helm-chart" />
  <Card title="Probes & drain" description="liveness / readiness / startup semantics and graceful-shutdown tuning." href="/docs/octopus/kubernetes/probes-and-drain" />
  <Card title="Discovery" description="EndpointSlice-based service discovery inside the cluster." href="/docs/octopus/kubernetes/discovery" />
  <Card title="TLS & cert-manager" description="Terminate TLS from Gateway listener Secrets, hot-reloaded as they rotate." href="/docs/octopus/kubernetes/tls-cert-manager" />
</Cards>

## Deployment topics that apply in-cluster

<Cards>
  <Card title="Scaling" description="gateway.workers, replica counts, the HPA, and shared state across replicas." href="/docs/octopus/deployment/scaling" />
  <Card title="Graceful shutdown" description="SIGTERM → readiness 503 → pre_stop_delay → drain, and sizing terminationGracePeriodSeconds." href="/docs/octopus/deployment/graceful-shutdown" />
  <Card title="Production checklist" description="TLS, auth, probes, resource limits, and metrics scraping before you go live." href="/docs/octopus/deployment/production-checklist" />
</Cards>