Past incidents
Elevated AI response latency
Resolved - latency returned to normal after our AI provider recovered. No messages were lost.
14:32 UTC
Identified - increased latency traced to an upstream model provider. Responses are slower but succeeding.
13:58 UTC
Investigating - we are seeing higher-than-normal response times for AI answers.
13:41 UTC
Knowledge base scans delayed
Resolved - the scan queue has fully drained and new scans start immediately.
09:20 UTC
Monitoring - additional workers deployed; the backlog is clearing.
08:05 UTC
Investigating - a spike in scan jobs caused a processing backlog. Chat and dashboard are unaffected.
07:40 UTC
Brief notifications outage
Resolved - the websocket server was restarted and real-time features are working again.
22:14 UTC
Investigating - real-time notifications and take-over are temporarily unavailable. Chats still send and receive.
22:02 UTC
Need to report an issue?
Not seeing your problem reflected here? Let our team know.