Our backend experienced a catastrophic failure during production due to a new feature release.The feature allowed users to bulk import data, but it caused the PostgreSQL server to reach its connection limit.As a result, all queries timed out, leading to the login service, user dashboard, and health checks being unresponsive.The downtime lasted for 47 minutes, marking the longest outage our startup has ever faced.