What is enterprise-kubernetes-delivery-platform?

Enterprise Kubernetes Delivery Platform is a template repository that shows how a production style Kubernetes deployment pipeline can be structured. It is explicitly not a real business application. The included Go service is kept small on purpose so the project can focus on deployment architecture instead of application logic. The core idea is to build a Docker image once and promote the same immutable image tag through dev, staging, and prod environments using Kustomize overlays, rather than rebuilding the service for each environment. Environment specific behavior, such as replica counts, log level, and feature flags, comes from ConfigMaps, Secrets, and overlay patches instead of hardcoded values in the Go code. ArgoCD handles GitOps reconciliation: dev syncs automatically, while staging and prod require a pull request that changes the target overlay's image tag before ArgoCD applies it. Rolling back a bad deployment is done by reverting that Git commit, and ArgoCD brings the cluster back to the previous state. The initial continuous integration workflow only validates the code and container: it formats and tests the Go service, builds the Docker image without pushing it, scans the Dockerfile with Hadolint, scans the built image with Trivy, renders the Kustomize overlays, and checks the Kubernetes manifests and GitHub Actions workflows with Checkov. Several other production concerns are documented as architecture rather than fully wired up yet, including policy as code with Kyverno, externalized secrets synced from AWS Secrets Manager, autoscaling with HPA and Karpenter, progressive canary delivery using Argo Rollouts and Istio, and observability with Prometheus and Grafana. Each of these areas ships with a helper script that prints out the intended operational flow, and the repository includes incident runbooks and day two operations documentation describing how the finished platform is meant to be run.