Previous incidents

Postmortem

The incident was caused by our database background optimization process not running aggressively enough. This allowed data fragments to accumulate across tables used for real-time cache loading.

As the number of fragments increased, queries that normally completed in milliseconds were forced to scan across many more fragments than necessary. This significantly increased query latency and led to exhaustion of available database connections, which impacted data ingestion and co...

The service degradation issue has been fully resolved. Our team has restored normal performance and will continue to monitor the system to ensure stability.

Thank you for your patience.