Top Ten Worst Repositories to Host on GitHub

Explore the challenges of hosting diverse projects on GitHub in this 43-minute conference talk. Discover how GitHub's infrastructure team tackles unusual interactions and extreme use cases, from repositories with over-eager CI systems to those with expensive Git processing requirements. Learn about innovative projects like Spokes and Lariat, developed to ensure reliable hosting of Git repositories. Gain insights into replication strategies, heuristics, caching mechanisms, and quota applications that maintain system reliability. Examine real-world examples of problematic repositories, including kubernetes/kubernetes, whosonfirst-data/whosonfirst-data-venue-us-ny, and Cocoa Pods/Specs. Understand the complexities of very big networks and the solutions implemented to address them. Conclude with valuable lessons on managing diverse project hosting challenges in large-scale version control systems.