Where to edit kube-controller config for cluster

Rancher
1

I’d like to edit some RKE cluster options to tweak HPA on my RKE/kubernetes cluster (kube controller manager) . Here’s how to do it for K3s for example ( https://rancher.com/docs/k3s/latest/en/installation/#customizing-components )

How do I do this in Rancher / RKE ? I tried setting the downscale setting to enable faster scale-down but it doesn’t seem to be applying (scale down still takes a long time). How can I identify what is actually applied? I followed this guide: https://rancher.com/blog/2018/2018-08-06-k8s-hpa-resource-custom-metrics/

services: 
  etcd: 
    backup_config: 
      enabled: true
      interval_hours: 12
      retention: 6
    creation: "12h"
    extra_args: 
      election-timeout: "5000"
      heartbeat-interval: "500"
    retention: "72h"
    snapshot: false
  kube-api: 
    always_pull_images: false
    pod_security_policy: false
    service_node_port_range: "30000-32767"
  kube-controller: 
    extra_args:  =================================> modified these args
      horizontal-pod-autoscaler-downscale-delay: "30s" 
      horizontal-pod-autoscaler-sync-period: "15s"
      horizontal-pod-autoscaler-upscale-delay: "30s"
2

I tracked down the docker container on the master node and dumped the args to the kube-controller-manager . Looks like the parameters are passed in BUT for sure, the scaler doesn’t scale down in 30 s as configured (in the absence of load) . This works in upstream k8s . I’ll see if I can investigate further. Any ideas, in the mean time i.e. how can I view what values the kube-controller-manager actually uses (something similar to a kubectl describe), regardless of what is passed to it?

kube-controller-manager --configure-cloud-routes=false --node-monitor-grace-period=40s --profiling=false --service-account-private-key-file=/etc/kubernetes/ssl/kube-service-account-token-key.pem --horizontal-pod-autoscaler-upscale-delay=30s --enable-hostpath-provisioner=false --leader-elect=true --pod-eviction-timeout=5m0s --cluster-cidr=10.42.0.0/16 --root-ca-file=/etc/kubernetes/ssl/kube-ca.pem --service-cluster-ip-range=10.43.0.0/16 --address=0.0.0.0 --allocate-node-cidrs=true --v=2 --allow-untagged-cloud=true --terminated-pod-gc-threshold=1000 --cloud-provider= --kubeconfig=/etc/kubernetes/ssl/kubecfg-kube-controller-manager.yaml --horizontal-pod-autoscaler-downscale-delay=30s --horizontal-pod-autoscaler-sync-period=15s --use-service-account-credentials=true