Dropbox Reduces Monorepo Size by 77% to Boost Developer Velocity
Dropbox engineers successfully reduced their server monorepo size from 87GB to 20GB, achieving a 77% reduction that significantly improved developer productivity. The massive repository, which houses most of the company's backend infrastructure, had grown to a point where cloning took over an hour, severely impacting onboarding and continuous integration pipelines. Additionally, the rapid growth rate posed a risk of hitting GitHub’s 100GB hard limit. The team identified that Git’s default delta compression heuristic, which relies on the last 16 characters of file paths, was inefficient for their internationalization files. This caused excessive storage usage rather than saving space. By addressing this compression issue, Dropbox cut clone times to under 15 minutes and mitigated operational risks. The article details the technical investigation into repository bloat, explaining how standard version control mechanisms can backfire at scale. This optimization ensures that as Dropbox continues to build AI-powered features like Dash, the engineering loop remains efficient, allowing developers to pull, build, test, and ship code changes without significant friction or delays associated with large data transfers.
Wire timeline
Dropbox Reduces Monorepo Size by 77% to Boost Developer Velocity
Dropbox engineers successfully reduced their server monorepo size from 87GB to 20GB, achieving a 77% reduction that significantly improved developer productivity. The massive repository, which houses most of the company's backend infrastructure, had grown to a point where cloning took over an hour, severely impacting onboarding and continuous integration pipelines. Additionally, the rapid growth rate posed a risk of hitting GitHub’s 100GB hard limit. The team identified that Git’s default delta compression heuristic, which relies on the last 16 characters of file paths, was inefficient for their internationalization files. This caused excessive storage usage rather than saving space. By addressing this compression issue, Dropbox cut clone times to under 15 minutes and mitigated operational risks. The article details the technical investigation into repository bloat, explaining how standard version control mechanisms can backfire at scale. This optimization ensures that as Dropbox continues to build AI-powered features like Dash, the engineering loop remains efficient, allowing developers to pull, build, test, and ship code changes without significant friction or delays associated with large data transfers.
Dropbox Tech Blog