Provision a Google Kubernetes Engine (GKE) Kubernetes Cluster


This guide will show you how to set up a GKE cluster on Google Cloud Platform, commonly referred to as GCP. When completed, you will have a fully functional Kubernetes cluster ready to deploy FusionAuth to as well as a PostgreSQL database using GCP’s Cloud SQL. The following method uses the default settings when provisioning the GKE cluster with the required resources and services. It is recommended that you consult Google Kubernetes Engine for full GKE documentation, including any custom changes needed for your situation.


  • Google Cloud Platform account with sufficient IAM permissions to create resources.
  • gcloud - Command Line tool used to manage resources in Google Cloud. For installation instructions, see Installing Cloud SDK.
  • kubectl - Command line tool that interacts with the Kubernetes API server and is useful for managing Kubernetes clusters. Before proceeding, follow the installation documentation that corresponds to your platform found here. This guide will be using version 1.22


The resulting GKE cluster will use a VPC-native cluster in the us-west1 region of which consists of three availability zones (us-west1-a, us-west1-b, and us-west1-c). You will provision a Cloud SQL PostgreSQL instance to satisfy installation requirements of FusionAuth.

GCP provides a number of configuration options designed to meet specific needs based on cluster availability and workload types. For this example, you will use the Standard mode of operation.

GKE Architecture

Project Setup

Having installed the Cloud SDK, authorize gcloud to access GCP using your Google credentials:

Authorize gcloud

gcloud auth login

Now create a new project used to organize all of your Google Cloud resources:

Create a project

gcloud projects create fusionauth-gke-example \
  --name="FusionAuth GKE example"

You will want to set the newly created project as your default project. When you create resources and enable APIs, they will be assigned to your default project:

Set the default project

gcloud config set project fusionauth-gke-example

Output of the default project selection command

Create in progress for [].
Waiting for [operations/cp.999999999156305912] to finish...done.
Enabling service [] on project [fusionauth-gke-example]...

Enable Billing

In order to proceed, you will need to enable and link billing for your project. To do this:

  1. Navigate to the GCP Console.
  2. From the navigation menu, select Billing.
  3. Verify that billing is enabled for your project. If it’s not, follow the prompts to link a billing to your project.

Enable Required APIs

Enable the Kubernetes Engine API. This will allow you to make a service request to the API to create your GKE cluster:

Enable Kubernetes Engine API

gcloud services enable

You will need to enable the Cloud SQL API in order to create a PostgreSQL database. Run the following command to do this:

Enable Cloud SQL API

gcloud services enable

In order for our GKE cluster to communicate with PostgreSQL and Elasticsearch on internal IP addresses, the Service Networking API must be enabled:

Enable Cloud SQL API

gcloud services enable \

Configure the Network

Allocate an IP address range for private communication on your VPC:

Create a private IP address range

gcloud compute addresses create fusionauth-private-range \
  --global \
  --purpose=VPC_PEERING \
  --addresses= \
  --prefix-length=16 \
  --description="Fusionauth private IP range for GKE and SQL" \
Created [

In order for GKE to communicate with Cloud SQL and Elasticsearch over a private network you need to create a private connection from your VPC network to the underlying service producer network.

Create a private connection

gcloud services vpc-peerings connect \ \
  --ranges=fusionauth-private-range \
  --network=default \

Create a GKE Cluster

With your project configured, billing enabled, and the Kubernetes Engine API enabled, you can proceed to create your GKE cluster.

To create a new cluster, execute the following.

Create GKE cluster

gcloud container clusters create fusionauth-cluster \
  --num-nodes=1 \
  --region=us-west1 \
  --enable-ip-alias \
  --cluster-version=1.21.4-gke.2300 \
  --cluster-ipv4-cidr= \
  • num-nodes - The number of nodes to be created in each zone. In this example, you specify the region of which consists of three zones. Therefore you will have a total of 3 nodes.
  • region - The region to create the cluster.
  • enable-ip-alias - Indicates to create a VPC-native cluster. This greatly simplifies network connectivity when communicating with the database by making pod IP addresses natively routable within the cluster’s VPC network.
  • cluster-version - The Kubernetes version to use. [optional]
  • cluster-ipv4-cidr - Used to create the subnet’s secondary IP address range for Pods. [optional]
  • service-ip-range - Used to create the subnet’s secondary IP address range for Services. [optional]

For more information on the create command, see gcloud container clusters create documentation.

Update Kubernetes Configuration

If the create operation completed successfully, the last thing it will do is update your local ~/.kube file. If that didn’t happen for whatever reason, gcloud provides the following to update your configuration and set the newly created cluster as the active context. This will let you use kubectl to access your cluster.

Get and update Kubeconfig

gcloud container clusters get-credentials fusionauth-cluster

Verify Cluster Configuration

Execute the list command to see GKE clusters that have been configured.

Get cluster information

gcloud container clusters list

fusionauth-cluster  us-west1  1.21.4-gke.2300  e2-medium     1.21.4-gke.2300  3          RUNNING

You now have a fully functional provisioned EKS cluster. For good measure, view the nodes that have been created. Use kubectl to make requests to the Kubernetes API Server.

Get list of nodes running on the cluster

kubectl get nodes -o wide

NAME                                                STATUS   ROLES    AGE   VERSION            INTERNAL-IP   EXTERNAL-IP      OS-IMAGE                             KERNEL-VERSION   CONTAINER-RUNTIME
gke-fusionauth-cluster-default-pool-2a2e7af5-nrrb   Ready    <none>   66m   v1.21.4-gke.2300   Container-Optimized OS from Google   5.4.120+         containerd://1.4.4
gke-fusionauth-cluster-default-pool-30c935b6-0mt4   Ready    <none>   66m   v1.21.4-gke.2300    Container-Optimized OS from Google   5.4.120+         containerd://1.4.4
gke-fusionauth-cluster-default-pool-431a5f55-rf11   Ready    <none>   66m   v1.21.4-gke.2300    Container-Optimized OS from Google   5.4.120+         containerd://1.4.4

Great! You have three nodes in a READY status. You can proceed to setting up a database.

Create a Database

Create a SQL Cloud PostgreSQL instance required for FusionAuth installation.

Create Cloud SQL for PostgreSQL instance

gcloud beta sql instances create fusionauth-test-db \
  --project=fusionauth-gke-example \
  --database-version=POSTGRES_12 \
  --tier=db-g1-small  \
  --region=us-west1 \
  --network=default \
  • project - The Id of the shared VPC service.
  • database-version - Database engine type and version. See FusionAuth supported databases here.
  • tier - Machine type for a shared-core instance.
  • region - The region to create the cluster.
  • network - Network in the current project that the instance will be part of.
  • no-assign-ip - Disables assignment of a public IP address.

For more information on the create command, see gcloud beta sql instances create documentation.

Configure the Default User

Google cloud SQL requires that you execute the following to configure the postgres user.

Set admin user password

gcloud sql users set-password postgres \
  --instance=fusionauth-test-db \

Verify Database Creation

Get list of Cloud SQL instances in the current project

gcloud sql instances list

fusionauth-test-db3  POSTGRES_12       us-west1-a  db-g1-small        -            RUNNABLE

Deploy Elasticsearch using Elastic Cloud

Google Cloud offers its Elasticsearch Service through Elastic Cloud. This section will guide you through setting up your account and deploying an Elasticsearch cluster.

From the navigation menu in the GCP console, click on Elasticsearch Service and then click the Enable button. Follow the instructions on the next screen to setup a new Elastic Cloud subscription.

Elasticsearch service enable subscription screen

After you have set up a subscription you will land on the GCP Elasticsearch Service Overview page. From here, click on the Manage Via Elastic button near the top of the window.

GCP Elasticsearch service screen

This will redirect you to the Elastic Cloud website. Login to Elastic Cloud using your Google account credentials.

After logging in, you will arrive at your Elastic Cloud dashboard. To begin creating a new Elasticsearch cluster, click on the Create deployment button.

Elastic Cloud dashboard screen

Input a name for your deployment and again click on Create deployment.

Create deployment screen

At this point, your deployment is now being created. You will be presented with deployment credentials on the next page. Download or save your credentials as instructed.

Save credentials screen

When your deployment creation process is complete, click on the Continue button. You will then be directed to your Elastic Cloud dashboard and will see your new deployment listed.

Click on the name of your deployment to manage it.

Elastic Cloud dashboard with deployment screen

From this dashboard, you have access to all of the necessary endpoint information you will need to connect to your deployment.

Elastic Cloud deployment dashboard screen

Under Applications, click on the Copy endpoint link next to Elasticsearch to copy the URL to your clipboard. You will need to save this URL for use when deploying FusionAuth to your GKE cluster.

Next Steps

You now are running all the necessary infrastructure to deploy a containerized application to GKE.

Next up, Deploy FusionAuth in Kubernetes.