Kubernetes Gateway API with Contour

This guide walks you through installing the Gateway API CRDs, deploying Contour using Helm with the Gateway API enabled, creating a GatewayClass, updating the Contour ConfigMap, and restarting the Contour deployment.

Prerequisites:

A Kubernetes cluster.
kubectl installed and configured to connect to your cluster.
Helm installed.

Steps

Install Gateway API CRDs

kubectl apply -f https://github.com/kubernetes-sigs/gateway-api/releases/download/v1.2.0/standard-install.yaml

This command applies the necessary Custom Resource Definitions (CRDs) for the Gateway API to your cluster.

Install Contour using Helm

helm upgrade --install contour contour \
  --repo https://charts.bitnami.com/bitnami \
  --set fullnameOverride=gateway \
  --namespace projectcontour --create-namespace \
  --set contour.ingressClass.create=true \
  --set contour.ingressClass.name=contour \
  --set gateway.enabled=true \
  --set gateway.installCRDs=true

Explanation of Flags:

helm upgrade --install contour contour: Installs or upgrades the Contour chart with the release name “contour.”
--repo https://charts.bitnami.com/bitnami: Specifies the Bitnami Helm chart repository.
--set fullnameOverride=gateway: Overrides the default resource naming to use “gateway” as the base name. This is important for consistency.
--namespace projectcontour --create-namespace: Creates the projectcontour namespace if it doesn’t exist and installs Contour into it. Best practice is to deploy Contour into its namespace.
--set contour.ingressClass.create=true: Creates an IngressClass resource.
--set contour.ingressClass.name=contour: Sets the name of the IngressClass to “contour”. Important if you are using Ingress objects alongside Gateway API objects.
--set gateway.enabled=true: Enables the Gateway API support in Contour.
--set gateway.installCRDs=true: Installs Gateway CRDs. Might not be necessary as you installed them previously, but including it doesn’t hurt.

Create a GatewayClass

kubectl apply -f - <<EOF
apiVersion: gateway.networking.k8s.io/v1beta1
kind: GatewayClass
metadata:
  name: contour
spec:
  controllerName: projectcontour.io/gateway-controller
EOF

This creates a GatewayClass named “contour” that tells Kubernetes to use Contour’s Gateway controller for handling Gateways of this class. The controllerName is crucial for linking the GatewayClass to the Contour controller.

Update the Contour ConfigMap

kubectl apply -f - <<EOF
apiVersion: v1
kind: ConfigMap
metadata:
  name: contour
  namespace: projectcontour
data:
  contour.yaml: |
    gateway:
      controllerName: projectcontour.io/gateway-controller
EOF

This step configures Contour to be the controller for the gateway. It’s very important to ensure Contour processes the Gateway API resources. It explicitly sets the controllerName within Contour’s configuration.

Restart Contour Deployment

kubectl -n projectcontour rollout restart deployment/gateway-contour

This restarts the Contour deployment to apply the configuration changes to the ConfigMap. This ensures that Contour picks up the new controllerName setting.

Verification

After completing these steps, you can verify the setup by:

Checking the Contour logs for any errors.
Create a Gateway resource and observe that Contour picks it up and configures the load balancer accordingly.
Creating an HTTPRoute resource to route traffic through the Gateway.

January 28, 2025