Case Study: Cursor reduces crashes and speeds bug fixes with Sentry

Cursor reduces OOM crashes by 80% with Sentry

Cursor, an AI-powered code editor, faced a significant challenge in maintaining stability for its desktop application, which engineers often left open for hours. Their client infrastructure team struggled with a manual, unscalable debugging process that required in-person intervention to diagnose crashes and errors. They turned to Sentry for its error monitoring, tracing, and Session Replay capabilities to gain visibility into issues across their multi-process Electron application.

By implementing Sentry, the team gained reliable error data and the ability to work asynchronously. They used breadcrumbs to reconstruct events leading to crashes and Session Replay for internal users to visually diagnose problems. Sentry provided the crucial signals needed to identify and fix root causes, such as a stack overflow bug. Furthermore, Cursor built automated tooling on top of Sentry's data, which helped them achieve an 80% decrease in out-of-memory crashes.

Cursor