On May 19 between 14:56 UTC and 15:34 UTC, many parts of our application’s user interface were not loading, preventing users from accessing the functionality. API’s weren’t affected. No data was lost.
Upon receiving error alerts from our monitoring systems, we identified the software change causing the issue and reverted the change. The fix was fully deployed at 15:34 UTC, restoring full functionality for all affected users.
The change we released relied on a feature of our frontend platform that was not available in all parts of our application. Our automated tests and the code review were not able to catch this issue prior to delivery.
To prevent similar issues in the future, we will extend our frontend platform to ensure the missing feature is available in all parts of our application.