Node Hello is a fun little exercise in automating the deployment of a Node.js webapp with an nginx reverse proxy using EKS and Helm.

Dependencies:

• awscli
• aws-iam-authenticator
• docker-for-mac
• helm
• kubectl

Local Development

1.) Intialize Helm

helm init

2.) Deploy the Helm Chart

cd node-hello && helm install -f values/development-values.yaml .

3.) Test

curl localhost:30000

Staging/Production

1.) Deploy the Cloudformation template

cd cfn && ./deploy.sh

2.) Provision the Cluster

export KUBECONFIG=~/kubeconfig-node-hello
kubectl apply -f ~/node-hello-aws-auth-cm.yaml

3.) Wait for Nodes to be Ready

kubectl get nodes --watch

NOTE: If you don't see nodes registering, try killing the running instances. The AutoScale group will replace them and the new ones should join the cluster immediately.

4.) Initialize Helm

helm init

5.) Patch Tiller for RBAC

kubectl create serviceaccount --namespace kube-system tiller
kubectl create clusterrolebinding tiller-cluster-rule --clusterrole=cluster-admin --serviceaccount=kube-system:tiller
kubectl patch deploy --namespace kube-system tiller-deploy -p '{"spec":{"template":{"spec":{"serviceAccount":"tiller"}}}}'

6.) Deploy Ingress Controller

helm install stable/nginx-ingress --name nginx-ingress --set rbac.create=true

7.) Deploy the Helm Chart

cd node-hello && helm install -f values/production-values.yaml --namespace=production .
cd node-hello && helm install -f values/staging-values.yaml --namespace=staging .

8.) Test

curl --header 'Host: node-hello.com' $(kubectl get svc nginx-ingress-controller -n default -o jsonpath="{.status.loadBalancer.ingress[*].hostname}")

curl --header 'Host: staging.node-hello.com' $(kubectl get svc nginx-ingress-controller -n default -o jsonpath="{.status.loadBalancer.ingress[*].hostname}")

Logging

Cluster-level logs can be access via the kubectl logs -f {POD_NAME} command. The official nginx image creates a symbolic link from /var/log/nginx/access.log to /dev/stdout , and creates another symbolic link from /var/log/nginx/error.log to /dev/stderr , overwriting the log files and causing logs to be sent to the relevant special device instead.

To Do

• Logs to Cloudwatch Using Fluentd
• Pipeline w/ Gitlab-CI
• Move Nodes to Spotinst
• ALB Ingress Controller?
• ExternalDNS?