A node in our production Kubernetes cluster became unresponsive, causing a brief disruption to some services while workloads were automatically rescheduled to healthy nodes. The duration of the disruption was about 8 minutes.