I've recently (last few months) had several instances where #AWS says they are passing all health checks. But the serial console is not available, and the node is not responding to anything at all. All endpoints are down in the load balancers.

I've had to take manual intervention to terminate these instances so that the auto-scaler can bring up a new one and K8s can move the workload. Anyone else seeing this? This is us-west-2 #awscloud