newrelic-k8s-guestbook

Kubernetes Guestbook application

New Relic Employees can find a complete workshop here (https://docs.google.com/document/d/1aXeWcQVcn-YvJK-MCCLnBNWA8ptzf2ZLVXVFDg2imn0/edit)

This repository contains all information to build a Kubernetes sample application. This work is based on previous work of Drew Decker (https://github.com/wreckedred/) and Clay Smith (https://github.com/smithclay/).

This is still a simple application, but it has some more services:

• Frontend service: Node.js app serving the UI
• Parser service: Node.js app responsible for parsing a message and sending it to RabbitMQ
• Worker service: Node.js app listening to RabbitMQ and pushing the message to Redis
• Redis service: For storing the message in Redis
• Queue service: RabbitMQ as message bus

Pre-requisites

You need a Kubernetes cluster to deploy this applicaton.

For Minikube: start minikube and make sure we can connect to Minikube's Docker daemon

minikube start
eval $(minikube docker-env) # Do this in every terminal window you are using

Create a Kubernetes secret with your New Relic license key

kubectl create secret generic guestbook-secret --from-literal=new_relic_license_key='<YOUR KEY HERE>'

Install the APM metadata injection

• Navigate to the k8s/metadata folder: kubectl apply -f .
• Wait until the metadata-setup job is completed: kubectl get pods

newrelic-metadata-injection-deployment-56dbf48c6-wkbnc   1/1     Running     0          17s
newrelic-metadata-setup-zw8sl                            0/1     Completed   0          18s

Install kube-state-metrics

curl -o kube-state-metrics-1.4.zip https://codeload.github.com/kubernetes/kube-state-metrics/zip/release-1.4 && unzip kube-state-metrics-1.4.zip && kubectl apply -f kube-state-metrics-release-1.4/kubernetes

Install the New Relic Kubernetes integration

• Navigate to the k8s/newrelic/ folder: kubectl apply -f .

Deploy RabbitMQ and Redis

• Navigate to the k8s/app/ folder:

kubectl apply -f rabbitmq.yaml
kubectl apply -f redis.yaml

• Check with kubectl get pods and wait until these services are running

Deploy the Node.js microservices (frontend, parser, worker)

• Navigate to the k8s/app/ folder:

kubectl apply -f frontend.yaml
kubectl apply -f parser.yaml
kubectl apply -f worker.yaml

• Check where frontend is running: kubectl describe service frontend

-> Look for 'NodePort 31811', we need this PORT

• kubectl cluster-info

-> Look for 'master IP: 192.168.64.3', we need this IP

Open IP:PORT in your browser

Debugging

watch kubectl logs -l tier=frontend --tail=20
watch kubectl logs -l tier=parser --tail=20
watch kubectl logs -l tier=worker --tail=20
watch kubectl logs -l tier=queue --tail=20

Build the Docker images

• If you want to make changes to any of the Docker images, navigate to the relevant folder under app/ and build the image: docker build -t nrlabs/newrelic-k8s-guestbook-frontend .

Clean-up

kubectl delete -f . # ATTENTION, this will delete everything from the yaml files in the current folder

Scale a deployment

kubectl scale --replicas=2 deployment/parser