I am looking forward to have an autoscaling feature on Kolla Ansible wallaby series openstack.

I am using Senlin to create cluster, gnocchi for metrics and aodh for alarm. 

However I am facing issue with aodh alarm state as it gets stuck in the state in which it is. 

I also found once the load is hiking, the gnocchi metrics get inconsistent. Due to this also the alarm state sticks and do not trigger alarm.

To solve this, i relied upon collectd. But collectd service does not push any metric from the Hypervisor to gnocchi.

My objective is that, in case of a metric reaching the threshold, alarm should automatically trigger and cluster should scale up or down based on the load. Once the scale up or down task is successfully completed, alarm should return to normal state automatically.

Does anybody have any experience with this use case or wanna propose any other solution?

Thanks