Learn how to install, configure, and visualize Kubernetes metrics with Prometheus and Grafana for effective monitoring.
💡 Introduction
Welcome to the world of DevOps! Today, we’ll walk through setting up Prometheus and Grafana on a Kind Kubernetes cluster using Helm. But before we dive into the technical steps, let’s understand what these tools are and why they matter.
We’ll be using AWS to create an EC2 instance, which will act as our main machine for setting up the cluster. If you’re new to AWS, Kubernetes, Prometheus, or Grafana, don’t worry—I’ll break it down for you.
What Are These Tools?
AWS (Amazon Web Services): A cloud platform that provides services like computing, storage, and networking.
Kind (Kubernetes IN Docker): A tool that lets you run a lightweight Kubernetes cluster on your local machine using Docker. It’s similar to Minikube but uses container nodes instead of virtual machines.
Prometheus: A powerful monitoring tool that scrapes metrics from Kubernetes nodes, pods, and applications. You can also instrument your applications using Prometheus client libraries.
Grafana: A visualization tool that takes data from Prometheus (or other sources like Loki) and displays it in interactive dashboards.
Project Workflow
Here’s a step-by-step breakdown of what we’ll be doing:
✅ Create an EC2 instance (t2.medium)
✅ Install Docker and Kind on the instance
✅ Create a Kubernetes cluster using Kind
✅ Install and configure kubectl (Kubernetes command-line tool)
✅ Apply Kubernetes manifests for deployment
✅ Install Helm (a package manager for Kubernetes)
✅ Deploy the kube-prometheus-stack using Helm
✅ Access Prometheus on port 9090
✅ Access Grafana on port 3000
💡 Pre-requisites
Before we dive into setting up Prometheus and Grafana, let’s make sure you have the necessary requirements in place.
✨ What You’ll Need:
- An AWS Account: Since we’ll be creating an EC2 instance to act as our main machine, you need access to AWS. If you don’t have an account yet, you can sign up at aws.amazon.com.
- Basic Understanding of Kubernetes and Docker: While I’ll walk you through each step, having some familiarity with Kubernetes (k8s) and Docker will make things easier. If you’re new to these technologies, here’s a quick breakdown:
- Docker: A tool for containerizing applications, allowing them to run consistently across different environments.
- Kubernetes: An orchestration system that manages containerized applications across multiple nodes.
Once you have these prerequisites ready, we’re all set to move forward with the setup! 🚀
Step 1: Creating an AWS EC2 Instance
- Log in to your AWS account and navigate to the EC2 Dashboard.
- Click Launch Instance and configure the instance as follows:
-
Name:
Prom-Graf-Project - OS: Ubuntu (latest version)
-
Instance Type:
t2.medium - Key-Pair: Select an existing key or create a new one
- Storage: Allocate at least 15GB
-
Name:
- Click Launch Instance.
Connect to the Instance
Once the instance is running, SSH into it using:
ssh -i /path/to/key-pair.pem ubuntu@<EC2-PUBLIC-IP>
Install Docker
sudo apt update
sudo apt install docker.io -y
sudo usermod -aG docker $USER && newgrp docker
Step 2: Installing Kind and Kubectl
Clone the GitHub repo containing required scripts:
git clone https://github.com/Pravesh-Sudha/k8s-kind-voting-app.git
cd k8s-kind-voting-app/kind-cluster
Install Kind
chmod +x install_kind.sh
./install_kind.sh
Install Kubectl
chmod +x install_kubectl.sh
./install_kubectl.sh
kind create cluster --config=config.yml
Deploy Kubernetes Manifests
cd ../k8s-specifications
kubectl apply -f .
Step 3: Installing Helm, Prometheus, and Grafana
Install Helm
curl -fsSL -o get_helm.sh https://raw.githubusercontent.com/helm/helm/main/scripts/get-helm-3
chmod 700 get_helm.sh
./get_helm.sh
Install Kube-Prometheus-Stack
helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
helm repo add stable https://charts.helm.sh/stable
helm repo update
kubectl create namespace monitoring
helm install kind-prometheus prometheus-community/kube-prometheus-stack --namespace monitoring \
--set prometheus.service.nodePort=30000 \
--set prometheus.service.type=NodePort \
--set grafana.service.nodePort=31000 \
--set grafana.service.type=NodePort
Verify Deployment
kubectl get svc -n monitoring
kubectl get namespace
Expose Prometheus and Grafana Services
kubectl port-forward svc/kind-prometheus-kube-prome-prometheus -n monitoring 9090:9090 --address=0.0.0.0 &
kubectl port-forward svc/kind-prometheus-grafana -n monitoring 3000:80 --address=0.0.0.0 &
-
Prometheus Dashboard:
http://<EC2-IP>:9090
-
Grafana Dashboard:
http://<EC2-IP>:3000
Step 4: Running Prometheus Queries
Open Prometheus Dashboard (http://<EC2-IP>:9090) and run the following queries:
CPU Usage (%)
sum (rate (container_cpu_usage_seconds_total{namespace="default"}[1m])) / sum (machine_cpu_cores) * 100
Memory Usage by Pod
sum (container_memory_usage_bytes{namespace="default"}) by (pod)
Network Data Received by Pod
sum(rate(container_network_receive_bytes_total{namespace="default"}[5m])) by (pod)
Network Data Transmitted by Pod
sum(rate(container_network_transmit_bytes_total{namespace="default"}[5m])) by (pod)
Generate Traffic
Expose the voting application:
kubectl port-forward service/vote 5000:5000 --address=0.0.0.0 &
Visit http://<EC2-IP>:5000, interact with the app (cast votes, refresh), then re-run queries in Prometheus.
Step 5: Visualizing Metrics in Grafana
- Open Grafana at
http://<EC2-IP>:31000 -
Login Credentials:
-
Username:
admin -
Password:
prom-operator
-
Username:
- Navigate to Dashboard → New → Import Dashboard
- Enter Dashboard ID:
15661 - Set the data source as Prometheus and click Import.
Your metrics are now visualized in Grafana! 🎉
Conclusion
With this setup, you now have a fully functional monitoring system for your Kubernetes cluster using Prometheus and Grafana. You can further:
✅ Add custom Grafana dashboards
✅ Enable Alertmanager for alerts
✅ Integrate Loki for log monitoring
By continuously monitoring your cluster, you ensure better performance, reliability, and efficiency. Happy monitoring! 🚀
Top comments (2)
Thanks for sharing. This post is extremely good
Anything for the community :)