Skip to content

Latest commit

 

History

History
109 lines (72 loc) · 3.48 KB

Ingress.md

File metadata and controls

109 lines (72 loc) · 3.48 KB
Raw

Ingress

Load Balancers are pretty great. But using the previous method, for each service we want to expose outside our cluster we will need to create a new load balancer.

An Ingress Controller is a way we can multiplex our load balancer. That is we can have one Load Balancer with rules on the back end that point to different services.

Change Back ngx service to ClusterIP

kubectl edit svc ngx

Change from LoadBalancer to ClusterIP. You'll have to remove the NodePort that was created. This will delete the Network Load Balancer.

Deploy Nginx Ingress Controller

(See latest version to download: here. )

kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.41.2/deploy/static/provider/aws/deploy.yaml

This will create resources in the ingress-nginx name space.

We've updated the configMap to make it compress the output for better performance of our applications. The configMap has this now:

data:
  use-gzip: "true"
  gzip-level: "7"
  enable-brotli: "true"
  brotli-level: "7"

Running:

kubectl get svc -n ingress-nginx

Will show the NLB configuration. Let's open to that webpage and see what it shows. You'll see a 503 Service Unavailable error. That's because its not pointing to anything.

503

Deploy Default Backend Ingress Rule

You can have a customized 404 page so when people hit a domain that is not in your rules they can go there. For our purposes we will create an app.

kubectl apply -f default-backend.yaml

Now we have a default service running when no ingress rules are specified. We can make this page say anything we want, just point it to a different service.

This yaml file creates a deployment of an application called kuard which just gives information about the server we are on. The new part in this yaml file is the end of a kind: Ingress. This is where we would put rules. Since we don't have any rules in here, it just becomes the default app to route to.

kuard

Update our nginx App with an Ingress Rule

We have a domain called castlerock.ai that we can use for this service. Let's create an ingress rule. The file in this directory called ngx-ing.yaml has the information below:

---
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  annotations:
    kubernetes.io/ingress.class: nginx
  name: ngx
spec:
  rules:
    - host: k8s.castlerock.ai
      http:
        paths:
          - path: "/"
            pathType: Prefix
            backend:
              service:
                name: ngx
                port: 
                  number: 80

The rule here shows that anytime the host is k8s.castlerock.ai it will route to our ngx service.

kubectl apply -f ngx-ing.yaml

Then we can see all our ingress rules in the default namespaces:

kubectl get ing

To get the nginx service we have to specify the host name. Since we don't have that mapped yet, we can spoof with curl.

curl -H "Host: k8s.castlerock.ai" a33115333837945aa9abd45ae3deca3f-3345cc15f67e0899.elb.us-west-2.amazonaws.com

Now our ingress controller is up. We just need to set our DNS to hit this load balancer to map to the services. Remember: An ingress controller is just a normal kubernetes service that listens to ingress events and then maps to those when it is hit from the outside. It allows us to multiplex our LoadBalancer!