API degraded

A background database optimization exposed an issue in connection handling, causing slower responses and intermittent API errors.

We restored normal service through a database failover and cleared the accumulated requests. The database optimization has since completed successfully.

To prevent recurrence, we have updated these background operations to release connections promptly, preventing the buildup that caused this incident. We have also added automated tests for this scenario.

All metrics have recovered. We continue to monitor.

Issue resolved and database cluster fully operational again. We are monitoring and checking the root cause.

API recovered as the database cluster triggered failover to new nodes.

A database index build operation is currently degrading API performance leading to slower responses and refused/times out connections. Our team is investigating.