Skip to content

Latest commit

 

History

History
56 lines (32 loc) · 3.31 KB

lb-ingress.md

File metadata and controls

56 lines (32 loc) · 3.31 KB

Load Balancer and Ingress

Ingress and LoadBalancer

LoadBalancer is a legacy way of getting some amount of traffic into your application.

Load Balancer

Actually does two separate things inside of your cluster.

A Load Balancer gives access to one set of pods only but in our case there are two sets of pods that we want to expose to outside world (multi-server and multi-client).

k8s is going to reach out your cloud provider and it is going to create a Load Balancer. You need a Load Balancer outside of your cluster in order to access the load balancer inside of your cluster.

Ingress

Exposes a set of services to the outside world.
There are several different implementations of ingress. We are going to use ingress-nginx not kubernetes-ingress that is a project from nginx company.

ingress-nginx vs kubernetes-ingress

Setup of ingress-nginx changes depending on your environment (local, GC, AWS, Azure etc). In this course, we are going to use local and GC.

Behind the Scenes of Ingress

We are going to create Ingress Config file which is going to be a set of routing rules, feed that file into kubectl that will create this ingress controller. The ingress controller is going to make something that accepts incoming traffic.

Ingress Config

In the ingress-nginx, the Ingress Controller and the thing that accepts incoming traffic are the same thing.

Ingress Controller

Behind the Scenes with Google Cloud

A Google Cloud Load Balancer is created and it sends traffic to a LoadBalancer service (the one we said that is not being used anymore) inside the cluster which is going to eventually get that traffic into the nginx Pod that gets created by our nginx controller. After that, it is up to the nginx Pod to eventually send that traffic off to the appropriate service inside our cluster.

Another deployment is setup inside your cluster, that is called default-backend pod, that is used for a series of health checks to essentially make sure that our cluster is working as expected. In the ideal world, you would replace this default-backend by your API.

Ingress Controller GC

Ingress Nginx Deployment

Commands to execute:

  • kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/master/deploy/static/mandatory.yaml
  • minikube addons enable ingress

The Ingress Config File

ingress-service.yaml

  • metadata/annotations - Higher level configurations around ingress object that is got created.
    • kubernetes.io/ingress.class: nginx - tells the k8s that we want to create an ingress controller based on nginx project
    • nginx.ingress.kubernetes.io/rewrite-target: / - how our copy of nginx behaves. After deciding sending that to Server instead of Client, it is going to do some rewrite on the request, for example, removing the /api part.
  • rules- are saying that there are two possible paths that we can match traffic to.