Periodic scaleUp/scaleDown not working in Monasca based autoscaling

[Rakesh H S]

Used the following template:
https://github.com/openstack/heat-templates/blob/master/hot/monasca/autoscaling.yaml

After deploying the template, when the cpu load of the two servers was above 50, an extra server was spawned successfully. However, the cpu load did not decrease, and all the three servers(included the newly added) had cpu load more than 50. I was expecting a new server to be spawned, but never happened.

This is because, when a state is moved from let's say OK to ALARM, monasca will send Alarm Notification only 'once' and assumes that the service which received this notification will handle things until the ALARM state is moved back to OK.

(unlike ceilometer, wherein notifications are sent continuously till the state changes back to OK)

[Doug Szumski (dszumski)]

It sounds like you need to set the period to 60 seconds on the Monasca Webhook notification. It will then ping heat every 60 seconds until the alarm transitions back.