A configuration deployment caused elevated 503 error rates affecting approximately 4% of API requests to the US-West region. The deployment was rolled back and the issue was fully resolved. No data was lost.

A queue processor memory leak caused webhook events to accumulate without delivery for approximately 40 minutes. Events were not lost and were delivered in order once the processor was restarted. Webhook retry logic prevented any permanent failures.

Network saturation at one edge ingestion node caused file upload times to increase by approximately 3x for users in certain geographic regions. Traffic was redistributed to healthy nodes.