Skip to content

Network Load Balancer With Health Check

Health check on target groups can be controlled with following annotations:

Annotation Function specifies the protocol used when performing health check on targets. specifies the port used when performing health check on targets. When using target-type: instance with a service of type NodePort, the healthcheck port can be set to traffic-port to automatically point to the correct port. specifies the HTTP path when performing health check on targets. specifies the interval(in seconds) between health check of an individual target. specifies the timeout(in seconds) during which no response from a target means a failed health check. specifies the HTTP or gRPC status code that should be expected when doing health checks against the specified health check path. specifies the consecutive health checks successes required before considering an unhealthy target healthy. specifies the consecutive health check failures required before considering a target unhealthy.

The Load Balancer Controller currently ignores the timeout configuration due to the limitations on the AWS NLB. The default timeout for TCP is 10s and HTTP is 6s.

Docker Images

Here is the Docker Image used in this tutorial: reyanshkharga/nodeapp:v1


reyanshkharga/nodeapp:v1 runs on port 5000 and has the following routes:

  • GET / Returns host info and app version
  • GET /health Returns health status of the app
  • GET /random Returns a randomly generated number between 1 and 10

Step 1: Create a Deployment

First, let's create a deployment as follows:

apiVersion: apps/v1
kind: Deployment
  name: my-deployment
  replicas: 2
      app: demo
        app: demo
      - name: nodeapp
        image: reyanshkharga/nodeapp:v1
          - containerPort: 5000

Apply the manifest to create the deployment:

kubectl apply -f my-deployment.yml

Verify deployment and pods:

# List deployments
kubectl get deployments

# List pods
kubectl get pods

Step 2: Create a LoadBalancer Service

Let's create a LoadBalancer service as follows:

apiVersion: v1
kind: Service
  name: nlb-service
  annotations: my-nlb external instance # Must specify this annotation internet-facing # Default is internal
    # Health Check http traffic-port /health '10' '2' # ignored '2' '2'
  type: LoadBalancer
    app: demo
    - port: 80
      targetPort: 5000

Apply the manifest to create the service:

kubectl apply -f my-service.yml

Verify service:

kubectl get svc

Note that we are offloading the reconciliation to AWS Load Balancer Controller using the external annotation.

Step 3: Verify AWS Resources in AWS Console

Visit the AWS console and verify the resources created by AWS Load Balancer Controller.

Pay close attention to the health check configuration of the target group that ingress created.

Note that the Load Balancer takes some time to become Active.

Also, verify that the NLB was created by AWS Load Balancer Controller. You can check the events in the logs as follows:

kubectl logs -f deploy/aws-load-balancer-controller -n aws-load-balancer-controller --all-containers=true

Step 4: Access App Via Network Load Balancer DNS

Once the load balancer is in Active state, you can hit the load balancer DNS and verify if everything is working properly.

Access the load balancer DNS by entering it in your browser. You can get the load balancer DNS either from the AWS console or the Ingress configuration.

Try accessing the following paths:

# Root path

# Health path

# Random generator path

Clean Up

Assuming your folder structure looks like the one below:

|-- manifests
│   |-- my-deployment.yml
│   |-- my-service.yml

Let's delete all the resources we created:

kubectl delete -f manifests/