We present the design, development, and evaluation of an automatic grading system that integrates Moodle with a Kubernetes cluster, aiming to be sufficiently general for use across a variety of courses. Experimentation in two large courses demonstrated the system's capability to support diverse programming assignments, handle substantial concurrent user loads, and enhance the educational experience. Future work will focus on system refinements, exploring additional integrations, and open-sourcing the project to benefit the wider educational community.