How to use Heat to auto scale across multiple datacenters
I am interested in understanding if I can get alarms on aggregated statistics using Heat. For example, I would want an alarm if an individual VM starts using more than 80% CPU, but for the whole group of VMs I want an alarm if average cpu_util goes over 40%. This is because I expect to have two geographic sites that can back each other up. So 40% or less should be the normal steady state so that if a site goes down the other site can handle the load by going up to 80%.
Also, once an alarm is raised does Heat give me the option to increase resources on a VM or only to create a new instance?
I have already looked at Heat docs on openstack.org but found it hard to get the answers for these questions there.