We use auto-scaling in Amazon (AWS) to scale-out new machines (or scale back) depending on CPU load. The machines that auto-scale out receive a known IP address from a pool of IPs. So I know all potential IPs. There is always 1 server running but if CPU load is heavy it would auto scale to 2 or 3 or up to 6 machines.
If I add each machine as a device in PRTG, I'll see critical warnings because the machines aren't always UP. Typically 1 or 2 will be up. They're only UP based on current CPU load. So I don't want to receive warnings/e-mails/texts once a scaled machine scales back down (turns off). I'm trying to figure out how to cleanly monitor machines that may or may not be present at a given moment: 2 machines may be up but all 6 could be up too for a while while the load is heavy.
I'm not sure if some sort of scripting or flag file could be used. Or maybe PRTG can be told to "pause" monitoring when a machine scales back down (off).
I'm sure others must have encountered this scenario while trying to monitor scaled instances.