Try Out the Cluster

After the cluster is created with a Pulumi update, there will be outputs with fields like the cluster’s kubeconfig file contents, and its cluster name for reference.

The full code for this stack is on GitHub.

The full code for this stack is on GitHub.

The full code for this stack is on GitHub.

Overview

We’ll explore how to:

Access the Cluster

In EKS, the account caller will be placed into the system:masters Kubernetes RBAC group by default. The kubeconfig generated will be specific to this primary cluster creator use-case, and it must be copied, and reconfigured to use with other IAM roles the caller assumes, as demonstrated in Configure Access Control.

As an Admin

Authentication

Authenticate as the admins role from the Identity stack.

$ aws sts assume-role --role-arn `pulumi stack output adminsIamRoleArn` --role-session-name k8s-admin
Kubeconfig Setup

To access your new Kubernetes cluster using kubectl, we need to setup the kubeconfig file, and export the environment variable for kubectl usage from the Cluster Configuration stack.

Setup the kubeconfig environment variable.

$ export KUBECONFIG=`pwd`/kubeconfig-admin.json

Get the Admins IAM Role ARN.

$ pulumi stack output adminsIamRoleArn
arn:aws:iam::000000000000:role/admins-eksClusterAdmin-0627674

Make a copy of the kubeconfig file that will be edited for the admins to use the adminsIamRoleArn output.

$ pulumi stack output kubeconfig > kubeconfig-admin.json

Edit kubeconfig-admin.json to use a role for authentication in the args of the aws-iam-authenticator, e.g.

...
"users": [
  {
    "name": "aws",
    "user": {
      "exec": {
        "apiVersion": "client.authentication.k8s.io/v1alpha1",
        "args": [
          "token",
          "-i",
          "k8s-aws-cluster-eksCluster-1ef1afe",
          "-r",
          "arn:aws:iam::000000000000:role/admins-eksClusterAdmin-0627674"
        ],
        "command": "aws-iam-authenticator"
      }
    }
  }
]

As a Developer

Authentication

Authenticate as the devs role from the Identity stack.

$ aws sts assume-role --role-arn `pulumi stack output devsIamRoleArn` --role-session-name k8s-devs
Kubeconfig Setup

To access your new Kubernetes cluster using kubectl, we need to setup the kubeconfig file, and export the environment variable for kubectl usage from the Cluster Configuration stack.

Setup the kubeconfig environment variable.

$ export KUBECONFIG=`pwd`/kubeconfig-devs.json

Get the Devs IAM Role ARN.

$ pulumi stack output devsIamRoleArn
arn:aws:iam::000000000000:role/devs-eksClusterDeveloper-e332028

Make a copy of the kubeconfig file that will be edited for the devs to use the devsIamRoleArn output.

$ pulumi stack output kubeconfig > kubeconfig-devs.json

Edit kubeconfig-devs.json to use a role for authentication in the args of the aws-iam-authenticator, e.g.

...
"users": [
  {
    "name": "aws",
    "user": {
      "exec": {
        "apiVersion": "client.authentication.k8s.io/v1alpha1",
        "args": [
          "token",
          "-i",
          "k8s-aws-cluster-eksCluster-1ef1afe",
          "-r",
          "arn:aws:iam::000000000000:role/devs-eksClusterDeveloper-e332028"
        ],
        "command": "aws-iam-authenticator"
      }
    }
  }
]

In AKS, the account caller will be placed into the system:masters Kubernetes RBAC group by default. Two kubeconfig files will be generated that will be specific to the admin and cluster user use-cases.

To configure the cluster for use with IAM roles, check out Configure Access Control.

Authentication

Authenticate as the ServicePrincipal from the Identity stack.

$ az login --service-principal --username $ARM_CLIENT_ID --password $ARM_CLIENT_SECRET --tenant $ARM_TENANT_ID

Admin Kubeconfig Setup

To access your new Kubernetes cluster using kubectl, we need to setup the kubeconfig file.

$ pulumi stack output kubeconfigAdmin > kubeconfig-admin.json
$ export KUBECONFIG=`pwd`/kubeconfig-admin.json

Developers Kubeconfig Setup

To access your new Kubernetes cluster using kubectl, we need to setup the kubeconfig file.

$ pulumi stack output kubeconfig > kubeconfig-devs.json
$ export KUBECONFIG=`pwd`/kubeconfig-devs.json

In GCP, the account caller will be placed into the system:masters Kubernetes RBAC group by default. The kubeconfig generated will be specific to this primary cluster creator use-case.

GCP authentication will use tokens to operate as Members such as Users or ServiceAccounts, and with certain permissions as detailed in Configure Access Control.

Admin Authentication

Authenticate as the admins ServiceAccount from the Identity stack.

$ pulumi stack output adminsIamServiceAccountSecret > k8s-admin-sa-key.json
$ gcloud auth activate-service-account --key-file k8s-admin-sa-key.json

Developer Authentication

Authenticate as the devs ServiceAccount from the Identity stack.

$ pulumi stack output devsIamServiceAccountSecret > k8s-devs-sa-key.json
$ gcloud auth activate-service-account --key-file k8s-devs-sa-key.json

Kubeconfig Setup

To access your new Kubernetes cluster using kubectl, we need to setup the kubeconfig file, and export the environment variable for kubectl usage.

$ pulumi stack output --show-secrets kubeconfig > kubeconfig.json
$ export KUBECONFIG=`pwd`/kubeconfig.json

Query the Cluster

Get cluster information.

$ kubectl version
$ kubectl cluster-info

Get the Nodes.

$ kubectl get nodes -o wide --show-labels

Get all Pods in the cluster, and show output attributes.

$ kubectl get pods --all-namespaces -o wide --show-labels

Get all Pods in the designated developer Namespace, and show output attributes.

$ kubectl get pods -n `pulumi stack output appsNamespaceName` -o wide --show-labels

Get the ConfigMaps of the kube-system Namespace.

$ kubectl get cm -n kube-system

Deploy a Workload

Imperatively deploy a NGINX Pod and public load-balanced service:

$ kubectl run --generator=run-pod/v1 nginx --image=nginx --port=80 --expose --service-overrides='{"spec":{"type":"LoadBalancer"}}'

After a few moments once it is deployed, visit the load balancer URL.

$ if ING_LB=$((kubectl get svc nginx -o template --template='{{(index .status.loadBalancer.ingress 0).hostname}}') 2>&1) ; then echo "http://$ING_LB"; else echo "LB is not ready yet."; fi

$ if ING_LB=$((kubectl get svc nginx -o template --template='{{(index .status.loadBalancer.ingress 0).ip}}') 2>&1) ; then echo "http://$ING_LB"; else echo "LB is not ready yet."; fi

$ if ING_LB=$((kubectl get svc nginx -o template --template='{{(index .status.loadBalancer.ingress 0).ip}}') 2>&1) ; then echo "http://$ING_LB"; else echo "LB is not ready yet."; fi

Delete the pod and service.

$ kubectl delete pod/nginx svc/nginx

Declaratively deploy a NGINX Pod and public load-balanced service:

import * as k8s from "@pulumi/kubernetes";

// Expose a k8s provider instance of the cluster.
const provider = new k8s.Provider("provider", {kubeconfig: kubeconfig });

// Create a NGINX Pod
const nginx = new k8s.core.v1.Pod(name,
    {
        metadata: {labels: {app: "nginx"}},
        spec: {
            containers: [
                {
                    name: name,
                    image: "nginx:latest",
                    ports: [{ name: "http", containerPort: 80 }]
                }
            ],
        }
    }, {provider: provider}
);

// Create a LoadBalancer Service for the NGINX Deployment
const service = new k8s.core.v1.Service(name,
    {
        metadata: {labels: {app: "nginx"}},
        spec: {
            type: "LoadBalancer",
            ports: [{ port: 80, targetPort: "http" }],
            selector: {app: "nginx"},
        },
    }, {provider: provider}
);
// Export the Service name and public LoadBalancer Endpoint
export const serviceName = service.metadata.name;
export const serviceHostname = service.status.loadBalancer.ingress[0].hostname;

After a few moments, visit the load balancer listed in the serviceHostname.

$ curl `pulumi stack output serviceHostname`

// Export the Service name and public LoadBalancer Endpoint
export const serviceName = service.metadata.name;
export const serviceIp = service.status.loadBalancer.ingress[0].ip;

After a few moments, visit the load balancer listed in the serviceIp.

$ curl `pulumi stack output serviceIp`

// Export the Service name and public LoadBalancer Endpoint
export const serviceName = service.metadata.name;
export const serviceIp = service.status.loadBalancer.ingress[0].ip;

After a few moments, visit the load balancer listed in the serviceIp.

$ curl `pulumi stack output serviceIp`

To tear down NGINX, delete its definition in the Pulumi program and run a Pulumi update.

Learn More

See the official Kubernetes Basics tutorial for more details.