Traffic Analyzer CI/CD + GitOps Project

Overview

This project demonstrates a production-style CI/CD pipeline using:

• Python (Flask) — application
• Docker — containerization
• Jenkins — CI (build and push)
• GitHub — source and configuration repositories
• Kubernetes — deployment platform
• ArgoCD — CD (GitOps)

The application is a Traffic Analyzer API that tracks incoming requests and exposes metrics.

---

Architecture

Developer (Git Push)
        ↓
GitHub (App Repo)
        ↓
Jenkins (CI Pipeline)
        ↓
DockerHub (Image Registry)
        ↓
GitHub (K8s Repo)
        ↓
ArgoCD (GitOps CD)
        ↓
Kubernetes Cluster
        ↓
Running Application

---

Project Structure

App Repository (traffic-analyzer)

traffic-analyzer/
├── app.py
├── requirements.txt
├── Dockerfile
├── Jenkinsfile

Kubernetes Repository (traffic-analyzer-k8s)

traffic-analyzer-k8s/
├── deployment.yml
├── service.yml

---

Application Details

Features

• Tracks total requests
• Tracks endpoint hits
• Tracks IP addresses

Endpoints

• / — Health check
• /stats — JSON analytics
• /metrics — Prometheus-compatible metrics

---

Docker

Build locally

docker build -t traffic-analyzer .
docker run -p 5000:5000 traffic-analyzer

---

Jenkins Pipeline

Stages

• Build Image
• Push Image to DockerHub
• Update Kubernetes Repository

CI Flow

git push → Jenkins builds image → pushes to DockerHub → updates K8s repo → ArgoCD deploys

---

Jenkins Credentials

ID	Purpose
dockerhub-creds	Push Docker images
github-creds	Update Kubernetes repository

---

Kubernetes Deployment

• Deployment with 2 replicas
• Uses versioned Docker images
• Service exposed via NodePort

---

ArgoCD (GitOps)

Responsibilities

• Watches the traffic-analyzer-k8s repository
• Applies changes automatically
• Keeps the cluster in sync with Git

Sync Policy

• Automatic sync enabled
• Prune enabled
• Self-heal enabled

---

End-to-End CI/CD Flow

1. Developer pushes code
2. Jenkins triggers (Poll SCM)
3. Jenkins builds and pushes the Docker image
4. Jenkins updates the Kubernetes repository
5. ArgoCD detects changes
6. ArgoCD syncs the deployment
7. Kubernetes performs a rolling update

---

Verification

Check running image

kubectl describe pod | grep Image:

Expected:

yogramming/traffic-analyzer:<latest-tag>

Watch rollout

kubectl get pods -w

---

Deployment Strategy

• Rolling updates (zero downtime)
• Old pods terminate after new pods are ready

---

Automation

• Jenkins uses Poll SCM (H/2 * * * *)
• Checks GitHub every ~2 minutes

---

Common Issues and Fixes

Issue	Cause	Fix
ImagePullBackOff	Image not pushed	Push image to DockerHub
Git push fails	Token permissions	Use PAT with write access
ArgoCD not deploying	Auto-sync disabled	Enable auto-sync
Jenkins not triggering	No trigger configured	Enable Poll SCM

---

Future Improvements

• Add Prometheus and Grafana dashboards
• Add Horizontal Pod Autoscaler (HPA)
• Add Ingress and domain routing
• Replace polling with GitHub webhooks
• Add staging and production environments

---

What This Project Demonstrates

• Real-world CI/CD pipeline
• GitOps workflow
• Kubernetes deployment strategies
• Docker image lifecycle
• Debugging real-world DevOps issues

---

Key Takeaways

• CI (Jenkins) builds artifacts
• CD (ArgoCD) deploys from Git
• Git is the single source of truth
• Kubernetes ensures safe deployments

---

Author

Built as a hands-on DevOps learning project focused on real-world workflows.

---

Final Note

This project reflects how modern cloud-native systems are built and deployed in production environments.