EKS with ArgoCD GitOps

In build — target Aug 31, 2026. The plan below is complete; real configs, command outputs, screenshots and the troubleshooting log land here as the build finishes.

EKS Terraform ArgoCD GitHub kubectl

What this project builds

A managed EKS cluster — two nodes across two availability zones — deployed by Terraform, with ArgoCD running in-cluster and watching a GitHub repository. Deployments happen by git push; rollbacks by git revert; the cluster continuously reconciles itself to what the repo declares.

How it will be built

EKS cluster and node group in Terraform, building on project 01's network patterns.
ArgoCD installed in-cluster, bootstrapped to watch an application repository.
Application manifests delivered the GitOps way: the repo is the source of truth, the cluster converges to it.
Add-ons installed by script; everything torn down nightly (~USD 5.50/day while alive).

Why it matters

I already run self-managed multi-master Kubernetes with automated failover in production, and I operate ArgoCD GitOps pipelines at work. This project is the managed-cloud counterpart — and knowing both sides of the self-managed vs managed trade-off, from experience, is exactly what the question "EKS or self-hosted?" is really asking.