What is the approach employed by kubelet to manage a currently running pod when the device within the pod is deemed unhealthy, as reported by the device plugin?

As I explore the code for the device plugin, I’ve observed that if a device is found to be in an unhealthy state, the device plugin will send a status update to kubelet. However, I’m uncertain about how kubelet manages a running pod when the device is declared unhealthy. Could you please direct me to the relevant code that handles this situation?