Ship it and Sleep

Modern CI/CD for Kubernetes-Native Teams.

This book targets senior developers and platform engineers who have a CI/CD pipeline that works until it does not. Never explains what a pipeline is. Never defines a container from scratch. The reader has shipped software to production. This book makes that process faster, safer, and recoverable at 3am.

Every chapter uses the same domain: a multi-service e-commerce platform. Product catalog, inventory, checkout, payments, and a frontend shell. Each service has its own repository. A shared infra repo holds all Kubernetes manifests. Pipeline examples, failure scenarios, and recovery procedures all refer to this platform. Abstract CI/CD problems become concrete and consistent across the book.

Four opinions run through every chapter:

GitHub Actions is the default CI platform. It has the broadest ecosystem, the most community-maintained actions, and the lowest barrier for a team that already lives in GitHub. GitLab CI is referenced where it handles a specific problem differently. Every pipeline example uses GitHub Actions syntax.

ArgoCD is the default GitOps engine. Flux is a legitimate alternative and gets an honest comparison, but ArgoCD's UI, app-of-apps model, and community adoption make it the practical default for teams adopting GitOps today. Every delivery example uses ArgoCD.

The GitOps repo is separate from the application repo. One repo per service for application code. One dedicated infra repo for all Kubernetes manifests, Helm values, and ArgoCD application definitions. This separation is enforced throughout the book. When a chapter touches both repos, it is explicit about which one.

Performance gates are non-negotiable. A deployment pipeline without a performance regression check is a pipeline that ships slowdowns silently. Every chapter that touches deployment includes or references a Locust gate. Locust is the performance testing tool for the entire book.

These positions are stated in chapter 1 and referenced when relevant. They are not hedged into "it depends" without a concrete condition and a decision rule.

This book was generated using AI assistance.