When creating deployment autoscale in AWS, I found the error below.

Error in autoscale nodes inside cluster Kubernetes. Failed "forbidden":

I0404 13:45:37.161486       1 scale_down.go:414] Node ip-192-1-3-138.us-east-2.compute.internal is not suitable for removal - utilization too big (0.975000)
I0404 13:45:37.161680       1 cluster.go:90] Fast evaluation: ip-192-1-3-52.us-east-2.compute.internal for removal
I0404 13:45:37.161797       1 static_autoscaler.go:381] ip-192-1-3-52.us-east-2.compute.internal is unneeded since 2019-04-04 13:45:06.622944386 +0000 UTC m=+30.741137175 duration 30.402283426s
I0404 13:45:37.161819       1 static_autoscaler.go:381] ip-192-1-3-132.us-east-2.compute.internal is unneeded since 2019-04-04 13:45:06.622944386 +0000 UTC m=+30.741137175 duration 30.402283426s
I0404 13:45:37.161829       1 static_autoscaler.go:392] Scale down status: unneededOnly=true lastScaleUpTime=2019-04-04 13:44:56.571655112 +0000 UTC m=+20.689847772 lastScaleDownDeleteTime=2019-04-04 13:44:56.571655909 +0000 UTC m=+20.689848562 lastScaleDownFailTime=2019-04-04 13:44:56.571656847 +0000 UTC m=+20.689850183 scaleDownForbidden=false isDeleteInProgress=false
I0404 13:45:37.747920       1 reflector.go:161] Listing and watching *v1.Job from k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:339
E0404 13:45:37.749860       1 reflector.go:126] k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:339: Failed to list *v1.Job: jobs.batch is forbidden: User "system:serviceaccount:kube-system:cluster-autoscaler" cannot list jobs.batch at the cluster scope
I0404 13:45:38.822647       1 reflector.go:161] Listing and watching *v1.Job from k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:339
E0404 13:45:38.825509       1 reflector.go:126] k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:339: Failed to list *v1.Job: jobs.batch is forbidden: User "system:serviceaccount:kube-system:cluster-autoscaler" cannot list jobs.batch at the cluster scope
I0404 13:45:39.922706       1 reflector.go:161] Listing and watching *v1.Job from k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:339
E0404 13:45:39.924756       1 reflector.go:126] k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:339: Failed to list *v1.Job: jobs.batch is forbidden: User "system:serviceaccount:kube-system:cluster-autoscaler" cannot list jobs.batch at the cluster scope

I used the deployment template (autoscaler):

https://github.com/kubernetes/autoscaler/blob/master/cluster-autoscaler/cloudprovider/aws/examples/cluster-autoscaler-one-asg.yaml

I solved this, adding rules needed of the Job on file cluster-autoscaler-one-asg.yaml (Kind: ClusterRole):

  - apiGroups: ["batch", "extensions"]
    resources: ["jobs"]
    verbs: ["get", "list", "watch", "patch"]