Getting Started

Steps follow the official Knative Quickstart. Commands assume a local Kubernetes cluster and a working kubectl.

Prerequisites

• A local Kubernetes cluster via kind or minikube.
• kubectl configured for that cluster.
• The kn CLI and its quickstart plugin from the Knative documentation.
• Container runtime (Docker or Podman) for the local cluster.

Install

The quickstart plugin provisions a local cluster with Serving, a networking layer, and DNS already wired up.

kn quickstart kind

For a manual install on an existing cluster, apply the Serving CRDs, then the core, then one networking layer, following the order in the Knative install docs. The sequence is CRDs, then core, then a networking layer such as Kourier, then DNS.

A first working setup

1. Confirm the Serving pods are running.

   kubectl get pods -n knative-serving

2. Deploy a service from a container image.

   kn service create hello \
     --image ghcr.io/knative/helloworld-go:latest \
     --port 8080 \
     --env TARGET=World

3. Knative creates a Configuration, an immutable Revision, and a Route, then prints the service URL once the Revision is Ready.

Verify it works

Curl the URL that kn service create printed. The first request after idle goes through the activator and incurs a cold start; the value should match the TARGET you set.

curl http://hello.default.<your-domain>

Replace <your-domain> with the host shown in the command output. After the ScaleToZeroGracePeriod of 30 seconds with no traffic (pkg/autoscaler/config/config.go:58), the Revision scales to zero, which you can watch with kubectl get pods. The next request triggers a cold start back up.

Where to go next

The Knative documentation covers production concerns this page does not: choosing and hardening a networking layer, configuring autoscaling targets and concurrency, traffic splitting across Revisions, TLS with cert-manager, and high-availability control-plane settings.