sept0r / teraki-grafana

Configure EKS cluster by terraform and setup Grafana application and Nginx ingress controller by helm

TERAKI Exercise

0. Prerequisites

Make shure you have the necessary permissions to create the EKS cluster.
You have installed kubectl, helm 3, awscli and aws-iam-authenticator.

1. Build EKS cluster using terraform

Go to the terraform directory.
Specify access_key and secret_key in the provider.tf or export environment variables:

export AWS_ACCESS_KEY_ID="AKIxxx"
export AWS_SECRET_ACCESS_KEY="xxx"

Change variables in the variables.tf file and run the following commands to setup EKS cluster:

terraform init
terraform apply

Generate the configuration for the kubectl and save it to the ~/.kube/config

terraform output kubeconfig > ~/.kube/config

2. Deploy Grafana application to the EKS cluster

Make sure you have the helm version 3.

helm version --short

Add and update the stable charts repo

helm repo add stable https://kubernetes-charts.storage.googleapis.com
helm repo update

You can change the default name and namespace to install the multiply Grafana applications.
Check the 4. Persistent storage and 5. Exposing Grafana to the internet for the details about parameters.

Create a new namespace:

kubectl create ns demo

To install the Grafana application from the helm chart, check the following parameters and run helm install.

helm install grafana-demo stable/grafana \
    --namespace demo \
    --set name=grafana-demo \
    --set service.type=ClusterIP \
    --set persistence.enabled=true \
    --set persistence.size=1Gi \
    --set ingress.enabled=true \
    --set "ingress.annotations.kubernetes\.io/ingress\.class=nginx"

Follow the instructions in the helm output to get the admin password and service IP address.

3. Exposing Grafana UI to the localhost

To expose the Grafana UI to the localhost use the following command:

kubectl port-forward service/grafana-demo 3000:80

The Grafana UI will be available on the port 3000. http://localhost:3000

4. Persistent storage

We have enabled the persistent storage for the Grafana deployment. By default it will be use a pvc persistence type.
To change the size of the persistent volume claim, change the persistence.size parameter in the helm install command.

5. Exposing Grafana to the internet

Install the Nginx ingress controller

To install the nginx-ingress from the heml chart use the following command. The controller.service.type will be set to LoadBalancer by default.

helm install nginx-demo stable/nginx-ingress --namespace demo

AWS LoadBalancer and Ingress resource

Get the AWS LoadBalancer DNS name:

kubectl get svc nginx-demo-nginx-ingress-controller

And patch the ingress resource:

kubectl patch ingress grafana-demo --type=json \
    -p='[{"op": "replace", "path": "/spec/rules/0/host", "value": "xxx-NNN.eu-central-1.elb.amazonaws.com"}]'

Grafana UI will be available by the LoadBalancer DNS.

6. SSL/TLS certificate

Get a new SSL/TLS certificate from the Let's Encript or other SSL provider and add it to the namespace

kubectl -n demo create secret tls frontend-ssl --key certificate.key --cert certificate-bundle.crt

Apply the new ingress resource

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  annotations:
    kubernetes.io/ingress.class: nginx
  name: grafana-demo
  namespace: demo
spec:
  tls:
  - hosts:
    - "xxx-NNN.eu-central-1.elb.amazonaws.com"
    secretName: frontend-ssl
  rules:
  - host: xxx-NNN.eu-central-1.elb.amazonaws.com
    http:
      paths:
      - backend:
          serviceName: grafana-demo
          servicePort: 80
        path: /