CI/CD workflow for Node.js application with Jenkins and Kubernetes

This repository provides information on how to set up a CI/CD workflow of a Node.js application hosted on Kubernetes.

Table of Contents

• Prerequisites
• Application Set-up Workflow
• File Structure

Prerequisites

• GitHub account.
• Docker Hub account.
• AWS account with (use Terraform to spin up servers):
  • Jenkins server (t2.micro)
  • SonarQube server (t2.medium)
  • Kubernetes Master server (t2.medium)
  • Kubernetes Worker One server (t2.micro)
  • Kubernetes Worker Two server (t2. micro)

Application Set-up Workflow

Step 1 — Setting up Jenkins server

• Create t2.micro instance in AWS and name for it. Example: jenkins-server
• Software to be installed on Jenkins machine:
  • Java
  • Git
  • Jenkins
  • Ansible
  • Terraform
• Once all the required software are installed Jenkins can be accessed via http://<jenkins-server-IP>:8080
• Login to Jenkins and install the below plugins:
  • Ansible plugin
  • NodeJS Plugin
  • SonarQube Scanner for Jenkins
  • Office 365 Connector

Step 2 — SonarQube server set-up and integration with Jenkins

• Create t2.medium instance in AWS and name for it. Example: sonar-qube-server
• Install SonarqQube Server software in sonar-qube-server and start sonar.
• Login to Jenkins and Integrate SonarQube server with Jenkins. (Jenkins -> Manage Jenkins -> Configure System -> SonarQube servers)

NOTE: SonarQube server can be access by http://<sonar-qube-server-IP>:9000

Step 3 — Set-up Kubernetes cluster

• Create 3 instances in AWS (use Terraform to spin up servers):
  • 1 t2.medium instance and name it. Example: K8-Master-Server
  • 2 t2.micro instances and name it. Example: K8-Worker-One, K8-Worker-Two
• Set-up Kubernetes cluster using Kubeadm.

Step 4 — Create sample Node.js application and push to GitHub

Step 5 — Create Pipeline job in Jenkins and Run it

Step 6 — Access your Node.js application on http://<Master/Worker-IP>:31200/

Note: In real world scenario the application would have been accessed via domain name. Ex: www.sampleapp.com which can be achieved using NGINX and Router 53 AWS service. However these tools were not used as they charge a lot for Elastic Load Balancer (ELB) and Router 53 in AWS.

Step 7 — Access monitoring tools

Prometheus http://<MasterIP>:30000/

Grafana http://<MasterIP>:32000/

File Structure

Explanations regarding structure:

• Dockerfile: contains instructions to build Docker image.
• Jenkinsfile: contains steps to run the Jenkins pipeline.
• MASTER-SERVER-FILES: Files in this folder are actually present in the Master Node. These files are placed here convenience purpose ONLY
  • nodejs-app-helm: helm chart for our Node.js project. All the required manifests files are placed inside the helm chart. Note: To demonstrate with and without referring inputs for manifest files from “values.yml” file the below actions are performed:
    • For “hpa.yml”, “service.yml” referred inputs from “values.yml”
    • For “deployment.yml”, ”namespace.yml” inputs are not referred from “values.yml”
  • prometheus-grafana: We can deploy Prometheus and Grafana easily using Helm charts. However to demonstrate the deployment without using Helm charts I have written all manifests file individually.
• package.json: records important metadata about a project which is required before publishing to NPM, and also defines functional attributes of a project that npm uses to install dependencies, run scripts, and identify the entry point to our package.
• playbooks: contains ansible playbooks which assists in deploying our Node.js application and monitoring tool in Kubernetes cluster.
  • hosts: stores information about remote nodes, which we need to manage.
  • kubernetes-helm-deployment.yml: playbook to deploy node.js application to Kubernetes application.
  • kubernetes-monitoring-deployment.yml: playbook to deploy Prometheus and Grafana to Kubernetes application.
• server.js: handles our application startup, routing and other functions.
• sonar-project.js: contains SonarQube server details, properties, etc.
• terraform-scripts: Infrastructure As Code (IAC) tool.
  • To demonstrate writing of terraform scripts with and without the use of “modules” the below actions are performed:
    • Used “modules” for
      • S3
      • DynamoDB
      • Terraform backend