01CI/CD · UNIVERSITY LAB
DevSecOps pipeline on AWS.
A university lab connecting Jenkins checks, ECR image publishing, and Argo CD delivery to Amazon EKS.
A commit starts the workflow. Source code and deployment configuration stay versioned in Git.
University / personal lab. The repository documents the implementation; this portfolio does not report live infrastructure health or production performance.
The problem.
A deployment involves more than building an image. I wanted to understand how source checks, image publishing, Kubernetes manifests, and delivery fit together—and where to look when one step fails.
The approach.
- Use Jenkins to coordinate builds, SonarQube analysis, and security scans with Trivy.
- Build frontend and backend container images and publish them to Amazon ECR.
- Keep Kubernetes application configuration in Git and use Argo CD for delivery to EKS.
- Practice service exposure, persistent storage, and monitoring with Prometheus and Grafana.
Decisions that mattered.
Separate CI from delivery
Jenkins builds and checks artifacts. Argo CD reconciles the configuration in Git with the Kubernetes cluster. That makes each tool’s responsibility easier to trace.
Check before publishing
Quality and security checks belong before delivery. A passing pipeline is useful evidence, but it does not establish that an application is secure in every environment.
Debug the boundary
For a failed rollout, start with the CI stage, then Argo CD sync, then Pod events and logs. An image pull failure and an application failure need different fixes.
Evidence & scope.
The linked repository describes the Jenkins, ECR, EKS, and Argo CD setup. The diagrams here explain the design; they are not screenshots of a running deployment. Pipeline runs have not been re-executed for this portfolio redesign.
Explore the repositoryWhat I took away.
Automation is easier to trust when every handoff has a clear owner and a useful place to look for evidence.
UP NEXT / 02