How we designed the GitLab Reference Architectures

GitLab introduced its Reference Architectures five years ago to provide scalable and elastic starting points for deploying GitLab at scale. The architectures were developed in partnership with the GitLab Test Platform and Support teams to address customer hurdles in deploying GitLab at scale. The goal was to provide best practices for deployment, scaling, and maintenance while reducing drift and increasing alignment. The Reference Architectures aimed to ensure performance, availability, scalability, cost-effectiveness, and maintainability. To design the architectures, GitLab gathered metrics from existing environments, defined a prototype architecture, built and tested it, and iteratively adjusted it based on test results. The process involved collecting data, defining a prototype, and testing and validating it through multiple cycles of iteration. The first validated Reference Architecture was published, and since then, the work has continued with additions such as various architecture sizes, step-by-step documentation, and cloud native hybrid variants. The Reference Architectures are constantly being improved and expanded through a comprehensive testing program that ensures they remain fit for purpose. The efforts have helped numerous customers and GitLab's own engineering teams, and the commitment to providing best-in-class guidance remains strong. Today, the Reference Architectures continue to evolve to meet the latest customer needs and GitLab code developments.