Between 11:01 and 13:41 CET, Kundo experienced two related periods of degraded service.
The first affected platform availability between 11:01 and 11:36, where slow or unavailable response times were observed. This was caused by a database migration running as part of a scheduled deployment that placed unexpected load on our database. The migration was stopped and the deployment was rolled back, restoring normal response times at 11:29.
The second period affected data delivery between 11:56 and 13:41, where users were not seeing updates to tickets and other changes. This was caused by a lag on our read replica database catching up after the migration was stopped. The issue was resolved by temporarily routing read traffic to the primary database while the replica recovered.
No data was lost. We have reviewed our deployment and migration routines and are implementing additional safeguards to prevent similar incidents in the future.