April 2021 Summaries

Sentry has introduced several features to help developers focus on what's truly important by silencing distractions and automating mundane tasks. The new filtering conditions allow users to filter issues by ownership, while the Review List provides a curated list of critical issues that need immediate attention, helping to drown out background noise. Additionally, Sentry now offers stack trace linking, which allows developers to jump directly into their codebase from within the platform. Furthermore, the integration with Jira and Azure DevOps has been improved to automate issue creation, allowing users to focus on bug fix rotation and making it more enjoyable.

In Sentry, a year-long effort was undertaken to convert its entire frontend React codebase from JavaScript to TypeScript, involving over a dozen team members, 1,100 files, and 95,000 lines of code. The decision was made due to the benefits of static analysis and type checking in preventing bugs, improving developer experience through editor integrations, reducing the need for API documentation, and leveraging an active community with rapid releases. The conversion was approached incrementally, with education, new code in TypeScript, and conversion phases, and faced challenges such as interoperability issues between TypeScript and React, incorrect library types, and maintaining motivation and energy throughout the process. After 18 months, the frontend codebase was successfully converted to TypeScript, resulting in improved protection against bugs, increased productivity, and greater confidence in shipping with fewer errors. The journey highlighted the importance of incremental conversion, staying current with TypeScript releases, building complex types incrementally, using TODO comments for future work, and introducing end-to-end type safety.

Using Feature Flags in Sentry to improve user experience and debug issues by safely controlling access to new features, using tags to identify specific users and experiments, and leveraging Discover to analyze errors and performance issues.