Cool-down is a confusing term in softlayer cloud. Group cool down serves a different purpose, it just waits before making any policy active and start again. It doesn’t monitor server(s), and never wait for changes inside the server, and hence doesn’t wait for server cool-down as the name suggests. Some would not even use cooldown time when initiating policies on specific date or time.
There is a need for creating policies twice when automating from both the end, viz, when server utilization increases, scale the server and when decreases de-scale the server. It can be achieved by creating two policies and two triggers – Given in the figure. Also the descaling depends upon termination policy (newest, oldest, or closet to next charge).
Rest Later, Good Bye!