A compute node did not automatically resume following an unscheduled restart, leading to instances being powered down and unable to boot. We are working to resolve the problem.
