Google Kubernetes Engine Deployment

ℹ️

Note (Gateway New Chart):

The Gateway new chart docs is now available here.

This page includes only Google Kubernetes Engine (GKE)-specific delta steps.

Review the Kubernetes Helm deployment page first, then apply the GKE changes in this guide.

Scope

This guide assumes that the baseline Helm deployment flow is complete, including:

  • Helm chart setup
  • base values.yaml preparation
  • installation and upgrade flow
  • Gateway admin and permission model

This page focuses on GKE workload identity and GKE-specific values.yaml changes.

Prerequisites

Complete all baseline prerequisites from the main Helm deployment page, and add:

GKE Identity Delta

Set the Gateway auth type to gcp and provide your GCP Access ID:

globalConfig:
  gatewayAuth:
    gatewayAccessId: <GCP Access ID>
    gatewayAccessType: gcp
  allowedAccessPermissions: {}

Set ServiceAccount annotations for workload identity:

serviceAccount:
  create: false
  serviceAccountName: <GKE ServiceAccount Name>
  annotations:
    iam.gke.io/gcp-service-account: <GCP Service Account>

nodeSelector:
  iam.gke.io/gke-metadata-server-enabled: "true"
ℹ️

Info:

For Autopilot clusters, omit nodeSelector. Autopilot rejects this selector because all nodes already use workload identity.

Validation Delta

After deployment, validate GKE workload identity integration:

  1. Confirm pod health:

    kubectl get pods -n <namespace>
  2. Confirm ServiceAccount annotations:

    kubectl get sa <GKE ServiceAccount Name> -n <namespace> -o yaml
  3. Validate Gateway login and management endpoint connectivity.

Related Tasks


Footer Section