Ngnix

Deployment Diagram

Nginx Deployment Diagram

Deploy Ngnix

Deployment

In the first step three Nginx instances are deployed to the Kubernetes Cluster. All three instances providing the same content, a simple "Hello World"

# Creates the deployment and rollout three nginx instances
kubectl apply -f k8s-apps/nginx/nginx-deployment.yaml

Next the nginx-service is deployed, this is a kubernetes resource to expose the pods to the internal network of the cluster.

Service

# Create a service which will route the traffic to one of the three nginx pods
kubectl apply -f k8s-apps/nginx/nginx-service.yaml

The pods are dynamically deployments, which means that pods can be created and destroyed at anytime. Consequently, the network configuration such as IP Addresses will change. Besides, all three pods are serving the same content and it is just relevant which pod will handle the request.

A Kubernetes Service solve this by giving as a stable DNS name and ensuring that the request it routed to one of the three pods.

Testing the Service

Services are just accessible within the Cluster, thus to test the configuration it is needed to create a Port-Forward from the client to the Kubernetes Cluster.

# Create a port-forwarding from local host to the service
kubectl port-forward service/my-ngnix-service 8080:8080

Opening http://192.168.60.10:8080 in a browser should show us the default Nginx start page.

Ingress

The goal of the last step is to expose the Nginx to the outside. Network traffic which is entering a network is called ingress, traffic leaving a network is called egress.

A Kubernetes Ingress consists of an Ingress Controller and an Ingress Resources. The controller is in a nutshell a http proxy, which will forward the traffic to the service/pod. In contrast is the resource the configuration of the controller behavior.

It is possible to use the same controller for several resources.

# Enable a Ingress controller
vagrant ssh -c "microk8s.enable ingress"

# Create a ingress
kubectl apply -f k8s-apps/nginx/nginx-ingress.yaml

The nginx is now reachable from outside the cluster, to test it open http://192.168.60.10/path-of-my-ingress in a Browser.

The pod deployment resource file nginx-deployment.yaml, specifics labels. These labels are useful to select a set of pods, e.g. for searching pods or linking resources together.
In the service resource file nginx-service.yaml the selector directive is used to select all pods with a specific set of labels. In this example the label app: nginx is used.
The name directive is used in the ingress resource nginx-service.yaml, to connect the ingress with the service.

Routing Network traffic

Nginx Network traffic

The nginx-ingress accept http request on port 80 from the client.
All traffic on port 80 is forwarded from the nginx-ingress to the nginx-service on Port 8080
The request which the nginx-service received on port 8080 is forwarded to exactly one nginx-pod on port 80