CALM Kubernetes Generator

Prerequisites

Node 20+
Docker desktop

A note about Docker Desktop on Mac

Version 4.34.0 of Docker Desktop has as a bug on Mac OS X that breaks Minikube. You should downgrade to the previous version. See https://github.com/docker/cli/issues/5412 for more information.

Installation

npm install
npm run build

Run the CLI

Either:

node dist/index.js

Or, after the link process has completed:

npx calm-k8s

Make it globally available

sudo npm link

Then it will be available everywhere as calm-k8s.

Example usage with Kubernetes

Minikube setup - Manually

Follow the steps for your platform here: https://minikube.sigs.k8s.io/docs/start/

You may also need to install kubectl on some Linux platforms.

Then start your cluster with the Calico CNI enabled.

minikube start --network-plugin=cni --cni=calico --kubernetes-version=1.30.0

Minikube setup - CALM + CLI

First run the CALM K8s CLI against the minikube templates. From the root of the project (same level as this README):

mkdir output
npx calm-k8s generate --templates templates/k8s-cluster-minikube --output output/minikube calm/instantiation.json

This will generate a script to set up the minikube cluster. Then run this script, again from the project root:

./output/minikube/initialise-cluster.sh

Generate & apply the resources

The templates/k8s-application directory contains templates to generate a set of Kubernetes resources for the application from the CALM instantiation. It also generates a kustomize script that makes applying the documents to your cluster easy.

From the project root, run the CLI against the k8s-application templates and output to the output/k8s-application directory:

npx calm-k8s generate --templates templates/k8s-application --output output/k8s-application calm/instantiation.json

You can now inspect the Kustomization with kubectl:

kubectl kustomize output/k8s-application

To apply the Kustomization: Note: -k instead of -f - this is to apply a Kustomization. Provide the directory when running with the -k argument.

kubectl apply -k output/k8s-application

Example output:

namespace/application created
service/application-svc created
service/db created
deployment.apps/application created
deployment.apps/postgres-database created
networkpolicy.networking.k8s.io/allow-egress-from-app-to-db created
networkpolicy.networking.k8s.io/allow-external-ingress-to-app created
networkpolicy.networking.k8s.io/allow-ingress-to-db-from-app created

You can verify the resources created with:

kubectl get all --namespace application

Example output:

NAME                              READY   STATUS    RESTARTS   AGE
pod/application-7bc585b64-b9nvj   1/1     Running   0          16s
pod/application-7bc585b64-dsz87   1/1     Running   0          16s
pod/application-7bc585b64-ltl47   1/1     Running   0          16s

NAME                      TYPE           CLUSTER-IP       EXTERNAL-IP   PORT(S)          AGE
service/application-svc   LoadBalancer   10.105.153.121   <pending>     8080:31005/TCP   17s

NAME                          READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/application   3/3     3            3           16s

NAME                                    DESIRED   CURRENT   READY   AGE
replicaset.apps/application-7bc585b64   3         3         3       16s

Access the service

The deployed application can be accessed via the load balancer using minikube tunnel. Details of how to use this: https://minikube.sigs.k8s.io/docs/handbook/accessing/#loadbalancer-access

Once the tunnel is active, the Swagger UI for the API of the application can be accessed locally at http://127.0.0.1:8080/q/swagger-ui/index.html.

jpgough / calm-k8s