Cluster autoscaling not enabled when flag is used

A cluster created with auto_scaling_enabled does have this feature. To the best of my knowledge this spans all currently available cluster templates. Known issues in 1.20.4 templates

 cat /etc/sysconfig/heat-params  | grep -i scal
AUTO_SCALING_ENABLED="True"
AUTOSCALER_TAG="v1.20.0-cern.0"

From the kubernetes side:

ka get po | grep scal
kube-system     kube-dns-autoscaler-58c879f5f6-xvgfl                      1/1     Running     0          4d20h

Checking the magnum src code, this should be enabled:

• By the upstream magnum installer when cern-magnum metachart is < 0.11.0 https://gitlab.cern.ch/kubernetes/magnum/-/blob/cern/train/magnum/drivers/common/templates/kubernetes/fragments/enable-auto-scaling.sh#L15

• By the cern metachart for cern-metachart versions >= 0.11.0| https://gitlab.cern.ch/kubernetes/magnum/-/blob/cern/train/magnum/drivers/common/templates/kubernetes/helm/cern-chart.sh#L149

Apparently the switch to enable this is not anywwhere in the magnum src code. The same issue might happen to other components https://gitlab.cern.ch/kubernetes/automation/releases/cern-magnum/-/blob/0.11.0/values.yaml#L454

• openstack-cloud-controller-manager
• coredns
• cluster-autoscaler

The original MR indeed does not include this activation in the helm chart: 5d835f0c