K8 services enable communication between components within and outside of K8 environment. For instance, it will allow users to connect with applications of the overall system such as frontend, backend, database & etc.

Services will also allow to loose couple our application between different microservices as shown below.

There are 3 types of services in K8.

NodePort - listens to a port on a node and then allows to access a port on a pod.
ClusterIP - Creation of a virtual IP inside the cluster that enables communication between backend and frontend services for example.
LoadBalancer - provisioning a load balancer for the application. Ex: distributing load across frontend tiers.

A service in K8 is an object (similar to node, pod or ReplicaSet) and one of its use case is to listen to port on the node and route incoming requests to the pods as shown below.

The above service is a NodePort service since it listens to a port on a node and allowing access to a port on a pod.

Let's have a closer look at the NodePort Service

It consist of 3 ports

TargetPort - port in the POD
Port - Port in the service
NodePort - port in the node (range: 30000 - 32767)

The services will be scaled across multi-node environments automatically without having to perform any additional configuration as shown below.

Creating and configuring a service on a cluster.

In this example, an Nginx app will be deployed and then HTTP access to that will be provided to that using a NordPort service. NodePort is useful when we need to give external access to the applications in a custom K8 cluster (using minikube, kubeadm & etc) - as opposed to using LoadBalanacers in AWS or Google Cloud.

Prerequisites

A fully functioning K8 cluster similar to this and configured with a master and a worker node. Up and running on a cloud environment (in this case Azure).
SSH access to the VMs
Admin access to enable/open ports in the VMs dashboard/firewall to allow HTTP traffic.

Step 1: Enabling HTTP access on the cloud environment. In this case install and enable Apache HTTP on the worker VM running on CentOS.

Note: a firewall isn't configured for this VM. Therefore it doesn't require any additional configuration to allow HTTP traffic to flow through.

Enabling Ports 80 and 30008 on worker VM.

Installing Apache server via SSH on the worker node.

yum install httpd && systemctl enable httpd && systemctl start httpd

After installing Apache server successfully, we should be able to access our VM using its public IP (20.37.11.77) on port 80 as shown below.

Step 2: Deploying the Nginx example app called myapp on the cluster via the following deployment script which consist of 2 replicas.

Note: Execute on master node

apiVersion: apps/v1
kind: Deployment
metadata:
  name: myapp-deployment
  labels:
    app: myapp
    type: front-end

spec:
  template:
    metadata:
      name: myapp-pod
      labels:
        app: myapp

    spec:
      containers:
        - name: nginx-container
          image: nginx

  replicas: 2

  selector:
    matchLabels:
      app: myapp

Creating the deployment

kubectl create -f deployment-definition.yml

Check the PODs

kubectl get pods -o wide

Step 3: Deploying the NordPort service that enables cURL and HTTP access to the above app that we just deployed.

Note: Execute on master node

apiVersion: v1
kind: Service
metadata:
  name: myapp-service

spec:
  type: NodePort
  ports:
    - targetPort: 80
      port: 80
      nodePort: 30008
  selector:
    app: myapp

The selector section will have the information of the POD name and the type (optional) - coming from the POD label section in its yaml file.
ports - within this, the only mandatory field is the port field. If we don’t specify nodePort, it'll automatically assign a value within the valid range - [30000 - 32767]. In this case we use 30008, which will be used for accessing the service via HTTP and cURL.
targetPort if not specified, then it is assumed to be the same as port.

Creating the service

kubectl create -f service-definition.yml

Checking the service deployment

kubectl get services

Step 4: Testing the service by accessing it via cURL and HTTP.

Using cURL - using the IP of the worker node and the port defined in the service yaml file above (30008).

curl 20.37.11.77:30008

Using HTTP - use any browser to access it