On 02/03/2015 04:17 PM, m.channappa.negalur at accenture.com wrote: > In heat auto scaling template we specify max_size as 3 , it means it can > launch upto 3 instances ?. I,e once high cpu load , crosses the > threshold level it will launch first vm (as scaling_adjustment : 1) and > if that load continues to remain same or increasing then the load will > be distributed on my second vm. If load is still keep on increasing > will it trigger 3^rd vm as max_size is 3 ..? is my assumptions correct. That is correct. Have a look at http://docs.openstack.org/hot-reference/content/OS__Heat__AutoScalingGroup.html for further parameters. At the moment you are missing a OS::Ceilometer::Alarm resource. This resource has to monitor the CPU load of the instances and trigger the scaling policy. Have a look at my simple example at https://blog.berendt.io/simple-auto-scaling-environment-with-heat/. Christian.