Server resources represent virtual machines / compute instances running in your cloud. AWS calls these EC2 instances while Azure calls them Azure Virtual Machines. Regardless of the cloud, PowerDown represents them as servers.
Servers are listed in the Resource List and look like this:
AWS EC2 will release public IP addresses when a server instance is stopped. When the resource is restarted, it will acquire a new public IP address. If you require a constant public IP, you should use an EC2 Elastic IP for the instance which will remain associated with a stopped instance.
PowerDown can be used for On-Demand, Reserved or Spot instances. However, reserved instances will not yield any savings with PowerDown because you have prepaid for the instance.
If a Spot instance is reclaimed by EC2, then it will be terminated and will be removed from PowerDown automatically. If you wish to use Spot instances with PowerDown, you should create an Autoscale group for the Spot instances and then attach your PowerDown schedule to the Autoscale group.
With reserved instances, you pay for a fixed term (1 to 3 years) regardless of whether the resource is running or stopped. Consequently, powering up or down the reserved instance will not impact the cloud provider bill.
On-demand instances are typically charged for by the second. Powering down on-demand instances will yield immediate savings.
If a server is a member of an Autoscale group it should not be powered up or down directly. Autoscale groups have a desired state policy and they will continually launch new instances if required to satisfy that desired state. Consequently, if you power down a server in an Autoscale group directly, the Autoscale group will immediately launch a new server to take it's place. You should manage such servers by powering up or down the Autoscale group.
PowerDown autoscale resources can override how many servers should be running when powered up and when powered down.