Introduction to K8s
This repository serves as an exploration of the basic usage of Kubernetes. It deploys a simple fastapi endpoint which greets the N-th user of the API. The visitor count is incremented using Redis. The whole application consists of two images: the fastapi server and redis. They are deployed in a single node--Control Node in the local kubernetes cluster thanks to Minikube.
Explain Like I'm 5 (ELI5)
In this repository, we will only tackle three things:
- Deployment
- Service
- Pods
Pods are the basic unit of kubernetes, it may contain a few docker images or containers. However, best practice suggest few containers or image as possible.
Services are used for communication between pods using a ClusterIP and for exposing pods outside the cluster using the NodePort
Deployments are used to define how many pods can be run, which group of pods to associate itself with, and the specifications of a container. Basically, it keeps the defined number of pods running. When one pod fails, it will spin up a new one with the same specification.
Try it out
Run your own kubernetes deployments locally.
Note: Make sure you have Docker, Kubectl, and Minikube installed.
- Clone the repository
git clone git@github.com:jrpespinas/intro-to-k8s.git- Start your minikube
minikube start- Deploy deployment definitions first to create instances of our images
kubectl create -f deployments/redis-deployment.yaml
kubectl create -f deployments/server-deployment.yaml- Deploy service definitions to expose the server to our localhost and to expose a ClusterIP to our redis
kubectl create -f deployments/redis-service.yaml
kubectl create -f deployments/server-service.yaml- Check if the pods are running
kubectl get pods,svc,deployment- If the status of the deployments are ready and if the pods are running, access the application through the given URL.
minikube service hello-service --url- You should see a similar output like this
{
"message": "Hello, visitor #1!"
}