The reboot process for AWS instances could be improved. Microsoft Azure does not have this problem, so AWS could consider making their instances more robust. You would not need to reboot your instances frequently to replace the hardware and stuff. They can look for a better approach or mechanism to improve in the future.
The concern is that you need to plan for the outage when you reboot an instance. You need to have a maintenance window where you can properly reboot the instance without affecting your application. When Amazon announces that you need to reboot an instance and are not ready, this becomes a problem.
