How GitLab uses Unicorn and unicorn-worker-killer

This post details the internal architecture of GitLab's web server (Unicorn) and its memory management (unicorn-worker-killer). It explains how Unicorn handles web requests with a multi-process architecture for better CPU utilization and fault tolerance, and how unicorn-worker-killer manages memory leaks in worker processes by automatically restarting them. This information is crucial for understanding the operational aspects and stability of self-managed GitLab installations.