This is where everything converges. Nine modules of knowledge -- Linux, networking, Docker, Kubernetes, Terraform, AWS, observability, SRE, streaming -- and now you deploy a real Rails application with the full stack. Not a tutorial app. Not a hello-world container. Your application, with your database, your background jobs, your uploads, your WebSocket connections, running on infrastructure you defined in code, monitored by dashboards you built, deployed by pipelines you wrote.
The capstone is the project that turns "Rails developer" into "Rails engineer who owns the full lifecycle." That distinction matters at the staff level. Staff engineers do not throw code over the wall to a DevOps team and hope it runs. They understand the platform, they own the deployment, they debug production issues across the full stack from the browser to the database query plan to the container resource limits.
Do not skip the chaos engineering exercise. I know it sounds reckless -- deliberately killing pods, injecting network latency, simulating disk failures on your own infrastructure. But breaking your own system teaches you more about resilience than any book or conference talk. You will discover that your health checks do not actually check health. You will discover that your "graceful shutdown" is not graceful. You will discover that your retry logic creates thundering herds. These are the lessons that only failure teaches, and it is better to learn them in a controlled experiment than at 3 AM during a real outage.
When you finish this capstone, you will have a portfolio piece that demonstrates end-to-end engineering competence. More importantly, you will have the confidence that comes from having built the thing, broken the thing, and fixed the thing. That confidence is not something you can fake in a staff-level interview, and it is not something you can get from reading documentation. You have to build it.