Our database infrastructure experienced elevated latency during a scheduled maintenance window, with response times spiking from ~150ms to ~13s. The platform remained accessible but slower than usual. Resolved automatically.