This guide walks you through how to configure the cluster-autoscaler on Amazon EKS.
edit templates/eks-autoscaler.yaml and specify the autoscaling group name in the command parameters.
For example:
command:
- ./cluster-autoscaler
- --v=4
- --stderrthreshold=info
- --cloud-provider=aws
- --skip-nodes-with-local-storage=false
- --nodes=2:8:EKS-myeks-DefaultNodeGroup-NodeGroup-1CXC7XVC6QY2K
This will specify the min as 2 and max to 8 on the EKS-myeks-DefaultNodeGroup-NodeGroup-1CXC7XVC6QY2K autoscaling group. You need to modify them for your own environment.
Attach extra IAM inline policy(template/asg-policy.json) to the EC2 IAM role.
$ kubectl apply -f templates/eks-autoscaler.yaml
serviceaccount "cluster-autoscaler" created
clusterrole.rbac.authorization.k8s.io "cluster-autoscaler" created
role.rbac.authorization.k8s.io "cluster-autoscaler" created
clusterrolebinding.rbac.authorization.k8s.io "cluster-autoscaler" created
rolebinding.rbac.authorization.k8s.io "cluster-autoscaler" created
deployment.extensions "cluster-autoscaler" created
$ kubectl logs -f deployment/cluster-autoscaler -n kube-system
And you should be able to see the log flowing like below
I0813 03:49:10.992856 1 flags.go:52] FLAG: --address=":8085"
I0813 03:49:10.992900 1 flags.go:52] FLAG: --alsologtostderr="false"
I0813 03:49:10.992905 1 flags.go:52] FLAG: --application-metrics-count-limit="100"
I0813 03:49:10.992908 1 flags.go:52] FLAG: --azure-container-registry-config=""
I0813 03:49:10.992919 1 flags.go:52] FLAG: --balance-similar-node-groups="false"
I0813 03:49:10.992922 1 flags.go:52] FLAG: --boot-id-file="/proc/sys/kernel/random/boot_id"
I0813 03:49:10.992925 1 flags.go:52] FLAG: --cloud-config=""
I0813 03:49:10.992929 1 flags.go:52] FLAG: --cloud-provider="aws"
I0813 03:49:10.992932 1 flags.go:52] FLAG: --cloud-provider-gce-lb-src-cidrs="130.211.0.0/22,209.85.152.0/22,209.85.204.0/22,35.191.0.0/16"
I0813 03:49:10.992938 1 flags.go:52] FLAG: --cluster-name=""
I0813 03:49:10.992941 1 flags.go:52] FLAG: --configmap=""
I0813 03:49:10.992949 1 flags.go:52] FLAG: --container-hints="/etc/cadvisor/container_hints.json"
I0813 03:49:10.992952 1 flags.go:52] FLAG: --containerd="unix:///var/run/containerd.sock"
I0813 03:49:10.992956 1 flags.go:52] FLAG: --cores-total="0:320000"
I0813 03:49:10.992959 1 flags.go:52] FLAG: --docker="unix:///var/run/docker.sock"
I0813 03:49:10.992963 1 flags.go:52] FLAG: --docker-env-metadata-whitelist=""
I0813 03:49:10.992966 1 flags.go:52] FLAG: --docker-only="false"
I0813 03:49:10.992969 1 flags.go:52] FLAG: --docker-root="/var/lib/docker"
I0813 03:49:10.992972 1 flags.go:52] FLAG: --docker-tls="false"
I0813 03:49:10.992975 1 flags.go:52] FLAG: --docker-tls-ca="ca.pem"
I0813 03:49:10.992978 1 flags.go:52] FLAG: --docker-tls-cert="cert.pem"
I0813 03:49:10.992981 1 flags.go:52] FLAG: --docker-tls-key="key.pem"
I0813 03:49:10.992984 1 flags.go:52] FLAG: --enable-load-reader="false"
I0813 03:49:10.992987 1 flags.go:52] FLAG: --estimator="binpacking"
I0813 03:49:10.992990 1 flags.go:52] FLAG: --event-storage-age-limit="default=0"
I0813 03:49:10.992993 1 flags.go:52] FLAG: --event-storage-event-limit="default=0"
I0813 03:49:10.992996 1 flags.go:52] FLAG: --expander="random"
I0813 03:49:10.993000 1 flags.go:52] FLAG: --expendable-pods-priority-cutoff="0"
I0813 03:49:10.993003 1 flags.go:52] FLAG: --gke-api-endpoint=""
I0813 03:49:10.993006 1 flags.go:52] FLAG: --global-housekeeping-interval="1m0s"
I0813 03:49:10.993011 1 flags.go:52] FLAG: --google-json-key=""
I0813 03:49:10.993014 1 flags.go:52] FLAG: --housekeeping-interval="10s"
I0813 03:49:10.993018 1 flags.go:52] FLAG: --httptest.serve=""
I0813 03:49:10.993023 1 flags.go:52] FLAG: --kubeconfig=""
I0813 03:49:10.993026 1 flags.go:52] FLAG: --kubernetes=""
I0813 03:49:10.993031 1 flags.go:52] FLAG: --leader-elect="true"
I0813 03:49:10.993040 1 flags.go:52] FLAG: --leader-elect-lease-duration="15s"
I0813 03:49:10.993045 1 flags.go:52] FLAG: --leader-elect-renew-deadline="10s"
I0813 03:49:10.993049 1 flags.go:52] FLAG: --leader-elect-resource-lock="endpoints"
I0813 03:49:10.993055 1 flags.go:52] FLAG: --leader-elect-retry-period="2s"
I0813 03:49:10.993058 1 flags.go:52] FLAG: --log-backtrace-at=":0"
I0813 03:49:10.993065 1 flags.go:52] FLAG: --log-cadvisor-usage="false"
I0813 03:49:10.993069 1 flags.go:52] FLAG: --log-dir=""
I0813 03:49:10.993074 1 flags.go:52] FLAG: --log-flush-frequency="5s"
I0813 03:49:10.993077 1 flags.go:52] FLAG: --logtostderr="true"
I0813 03:49:10.993081 1 flags.go:52] FLAG: --machine-id-file="/etc/machine-id,/var/lib/dbus/machine-id"
I0813 03:49:10.993087 1 flags.go:52] FLAG: --max-autoprovisioned-node-group-count="15"
I0813 03:49:10.993090 1 flags.go:52] FLAG: --max-empty-bulk-delete="10"
I0813 03:49:10.993095 1 flags.go:52] FLAG: --max-failing-time="15m0s"
I0813 03:49:10.993098 1 flags.go:52] FLAG: --max-graceful-termination-sec="600"
I0813 03:49:10.993104 1 flags.go:52] FLAG: --max-inactivity="10m0s"
I0813 03:49:10.993107 1 flags.go:52] FLAG: --max-node-provision-time="15m0s"
I0813 03:49:10.993110 1 flags.go:52] FLAG: --max-nodes-total="0"
I0813 03:49:10.993113 1 flags.go:52] FLAG: --max-total-unready-percentage="45"
I0813 03:49:10.993117 1 flags.go:52] FLAG: --memory-total="0:6400000"
I0813 03:49:10.993120 1 flags.go:52] FLAG: --min-replica-count="0"
I0813 03:49:10.993125 1 flags.go:52] FLAG: --namespace="kube-system"
I0813 03:49:10.993128 1 flags.go:52] FLAG: --node-autoprovisioning-enabled="false"
I0813 03:49:10.993132 1 flags.go:52] FLAG: --node-group-auto-discovery="[]"
I0813 03:49:10.993135 1 flags.go:52] FLAG: --nodes="[2:8:EKS-myeks-DefaultNodeGroup-NodeGroup-1CXC7XVC6QY2K]"
I0813 03:49:10.993139 1 flags.go:52] FLAG: --ok-total-unready-count="3"
I0813 03:49:10.993142 1 flags.go:52] FLAG: --regional="false"
I0813 03:49:10.993147 1 flags.go:52] FLAG: --scale-down-candidates-pool-min-count="50"
I0813 03:49:10.993150 1 flags.go:52] FLAG: --scale-down-candidates-pool-ratio="0.1"
I0813 03:49:10.993156 1 flags.go:52] FLAG: --scale-down-delay-after-add="10m0s"
I0813 03:49:10.993159 1 flags.go:52] FLAG: --scale-down-delay-after-delete="10s"
I0813 03:49:10.993163 1 flags.go:52] FLAG: --scale-down-delay-after-failure="3m0s"
I0813 03:49:10.993166 1 flags.go:52] FLAG: --scale-down-enabled="true"
I0813 03:49:10.993169 1 flags.go:52] FLAG: --scale-down-non-empty-candidates-count="30"
I0813 03:49:10.993172 1 flags.go:52] FLAG: --scale-down-unneeded-time="10m0s"
I0813 03:49:10.993175 1 flags.go:52] FLAG: --scale-down-unready-time="20m0s"
I0813 03:49:10.993178 1 flags.go:52] FLAG: --scale-down-utilization-threshold="0.5"
I0813 03:49:10.993181 1 flags.go:52] FLAG: --scan-interval="10s"
I0813 03:49:10.993184 1 flags.go:52] FLAG: --skip-nodes-with-local-storage="false"
I0813 03:49:10.993187 1 flags.go:52] FLAG: --skip-nodes-with-system-pods="true"
I0813 03:49:10.993190 1 flags.go:52] FLAG: --stderrthreshold="0"
I0813 03:49:10.993193 1 flags.go:52] FLAG: --storage-driver-buffer-duration="1m0s"
I0813 03:49:10.993196 1 flags.go:52] FLAG: --storage-driver-db="cadvisor"
I0813 03:49:10.993199 1 flags.go:52] FLAG: --storage-driver-host="localhost:8086"
I0813 03:49:10.993202 1 flags.go:52] FLAG: --storage-driver-password="root"
I0813 03:49:10.993205 1 flags.go:52] FLAG: --storage-driver-secure="false"
I0813 03:49:10.993208 1 flags.go:52] FLAG: --storage-driver-table="stats"
I0813 03:49:10.993211 1 flags.go:52] FLAG: --storage-driver-user="root"
I0813 03:49:10.993213 1 flags.go:52] FLAG: --test.bench=""
I0813 03:49:10.993216 1 flags.go:52] FLAG: --test.benchmem="false"
I0813 03:49:10.993219 1 flags.go:52] FLAG: --test.benchtime="1s"
I0813 03:49:10.993222 1 flags.go:52] FLAG: --test.blockprofile=""
I0813 03:49:10.993225 1 flags.go:52] FLAG: --test.blockprofilerate="1"
I0813 03:49:10.993228 1 flags.go:52] FLAG: --test.count="1"
I0813 03:49:10.993231 1 flags.go:52] FLAG: --test.coverprofile=""
I0813 03:49:10.993234 1 flags.go:52] FLAG: --test.cpu=""
I0813 03:49:10.993237 1 flags.go:52] FLAG: --test.cpuprofile=""
I0813 03:49:10.993243 1 flags.go:52] FLAG: --test.memprofile=""
I0813 03:49:10.993246 1 flags.go:52] FLAG: --test.memprofilerate="0"
I0813 03:49:10.993249 1 flags.go:52] FLAG: --test.mutexprofile=""
I0813 03:49:10.993252 1 flags.go:52] FLAG: --test.mutexprofilefraction="1"
I0813 03:49:10.993255 1 flags.go:52] FLAG: --test.outputdir=""
I0813 03:49:10.993257 1 flags.go:52] FLAG: --test.parallel="2"
I0813 03:49:10.993260 1 flags.go:52] FLAG: --test.run=""
I0813 03:49:10.993263 1 flags.go:52] FLAG: --test.short="false"
I0813 03:49:10.993266 1 flags.go:52] FLAG: --test.timeout="0s"
I0813 03:49:10.993269 1 flags.go:52] FLAG: --test.trace=""
I0813 03:49:10.993272 1 flags.go:52] FLAG: --test.v="false"
I0813 03:49:10.993275 1 flags.go:52] FLAG: --v="4"
I0813 03:49:10.993278 1 flags.go:52] FLAG: --version="false"
I0813 03:49:10.993284 1 flags.go:52] FLAG: --vmodule=""
I0813 03:49:10.993287 1 flags.go:52] FLAG: --write-status-configmap="true"
I0813 03:49:10.993293 1 main.go:298] Cluster Autoscaler 1.2.2
I0813 03:49:11.107635 1 leaderelection.go:175] attempting to acquire leader lease kube-system/cluster-autoscaler...
I0813 03:49:11.112799 1 leaderelection.go:243] lock is held by cluster-autoscaler-5c4c96764b-2q25v and has not yet expired
I0813 03:49:11.112815 1 leaderelection.go:180] failed to acquire lease kube-system/cluster-autoscaler
I0813 03:49:13.552811 1 leaderelection.go:243] lock is held by cluster-autoscaler-5c4c96764b-2q25v and has not yet expired
I0813 03:49:13.552935 1 leaderelection.go:180] failed to acquire lease kube-system/cluster-autoscaler
I0813 03:49:16.484256 1 leaderelection.go:243] lock is held by cluster-autoscaler-5c4c96764b-2q25v and has not yet expired
I0813 03:49:16.484275 1 leaderelection.go:180] failed to acquire lease kube-system/cluster-autoscaler
I0813 03:49:19.483446 1 leaderelection.go:243] lock is held by cluster-autoscaler-5c4c96764b-2q25v and has not yet expired
I0813 03:49:19.483467 1 leaderelection.go:180] failed to acquire lease kube-system/cluster-autoscaler
I0813 03:49:22.010573 1 leaderelection.go:243] lock is held by cluster-autoscaler-5c4c96764b-2q25v and has not yet expired
I0813 03:49:22.010598 1 leaderelection.go:180] failed to acquire lease kube-system/cluster-autoscaler
I0813 03:49:25.331903 1 leaderelection.go:243] lock is held by cluster-autoscaler-5c4c96764b-2q25v and has not yet expired
I0813 03:49:25.331916 1 leaderelection.go:180] failed to acquire lease kube-system/cluster-autoscaler
I0813 03:49:29.426059 1 leaderelection.go:184] successfully acquired lease kube-system/cluster-autoscaler
I0813 03:49:29.486473 1 predicates.go:125] Using predicate PodFitsResources
I0813 03:49:29.486568 1 predicates.go:125] Using predicate GeneralPredicates
I0813 03:49:29.486591 1 predicates.go:125] Using predicate PodToleratesNodeTaints
I0813 03:49:29.486611 1 predicates.go:125] Using predicate CheckNodeMemoryPressure
I0813 03:49:29.486629 1 predicates.go:125] Using predicate ready
I0813 03:49:29.486647 1 predicates.go:125] Using predicate CheckNodeDiskPressure
I0813 03:49:29.486761 1 predicates.go:125] Using predicate CheckVolumeBinding
I0813 03:49:29.486794 1 predicates.go:125] Using predicate MaxAzureDiskVolumeCount
I0813 03:49:29.486814 1 predicates.go:125] Using predicate NoDiskConflict
I0813 03:49:29.486833 1 predicates.go:125] Using predicate CheckNodeCondition
I0813 03:49:29.486851 1 predicates.go:125] Using predicate MaxGCEPDVolumeCount
I0813 03:49:29.486869 1 predicates.go:125] Using predicate MatchInterPodAffinity
I0813 03:49:29.486887 1 predicates.go:125] Using predicate MaxEBSVolumeCount
I0813 03:49:29.486974 1 predicates.go:125] Using predicate NoVolumeZoneConflict
I0813 03:49:29.487448 1 reflector.go:202] Starting reflector *v1beta1.DaemonSet (1h0m0s) from k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:293
I0813 03:49:29.487499 1 reflector.go:240] Listing and watching *v1beta1.DaemonSet from k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:293
I0813 03:49:29.488145 1 reflector.go:202] Starting reflector *v1beta1.StatefulSet (0s) from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.488186 1 reflector.go:240] Listing and watching *v1beta1.StatefulSet from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.488389 1 factory.go:33] Event(v1.ObjectReference{Kind:"Endpoints", Namespace:"kube-system", Name:"cluster-autoscaler", UID:"e9498bdb-9e9c-11e8-b9d7-06a77e9e2ace", APIVersion:"v1", ResourceVersion:"2904872", FieldPath:""}): type: 'Normal' reason: 'LeaderElection' cluster-autoscaler-5c4c96764b-c4mnk became leader
I0813 03:49:29.488931 1 reflector.go:202] Starting reflector *v1.Pod (1h0m0s) from k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:149
I0813 03:49:29.488958 1 reflector.go:240] Listing and watching *v1.Pod from k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:149
I0813 03:49:29.489405 1 reflector.go:202] Starting reflector *v1.PersistentVolumeClaim (0s) from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.489421 1 reflector.go:240] Listing and watching *v1.PersistentVolumeClaim from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.489896 1 reflector.go:202] Starting reflector *v1.Service (0s) from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.489914 1 reflector.go:240] Listing and watching *v1.Service from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.490436 1 reflector.go:202] Starting reflector *v1beta1.PodDisruptionBudget (0s) from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.490454 1 reflector.go:240] Listing and watching *v1beta1.PodDisruptionBudget from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.583446 1 reflector.go:202] Starting reflector *v1.Pod (1h0m0s) from k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:174
I0813 03:49:29.583493 1 reflector.go:240] Listing and watching *v1.Pod from k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:174
I0813 03:49:29.583761 1 reflector.go:202] Starting reflector *v1.Pod (0s) from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.583800 1 reflector.go:240] Listing and watching *v1.Pod from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.583807 1 reflector.go:202] Starting reflector *v1.Node (1h0m0s) from k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:212
I0813 03:49:29.583838 1 reflector.go:240] Listing and watching *v1.Node from k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:212
I0813 03:49:29.583972 1 reflector.go:202] Starting reflector *v1.Node (1h0m0s) from k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:239
I0813 03:49:29.583985 1 reflector.go:240] Listing and watching *v1.Node from k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:239
I0813 03:49:29.584115 1 reflector.go:202] Starting reflector *v1beta1.PodDisruptionBudget (1h0m0s) from k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:266
I0813 03:49:29.584128 1 reflector.go:240] Listing and watching *v1beta1.PodDisruptionBudget from k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes/listers.go:266
I0813 03:49:29.584186 1 reflector.go:202] Starting reflector *v1.Node (0s) from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.584211 1 reflector.go:240] Listing and watching *v1.Node from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.584358 1 reflector.go:202] Starting reflector *v1.PersistentVolume (0s) from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.584367 1 reflector.go:240] Listing and watching *v1.PersistentVolume from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.584544 1 reflector.go:202] Starting reflector *v1.StorageClass (0s) from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.584579 1 reflector.go:240] Listing and watching *v1.StorageClass from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.584897 1 reflector.go:202] Starting reflector *v1.ReplicationController (0s) from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.584933 1 reflector.go:240] Listing and watching *v1.ReplicationController from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.585221 1 reflector.go:202] Starting reflector *v1beta1.ReplicaSet (0s) from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.585258 1 reflector.go:240] Listing and watching *v1beta1.ReplicaSet from k8s.io/autoscaler/cluster-autoscaler/vendor/k8s.io/client-go/informers/factory.go:87
I0813 03:49:29.883414 1 request.go:481] Throttling request took 298.392935ms, request: GET:https://10.100.0.1:443/api/v1/replicationcontrollers?limit=500&resourceVersion=0
I0813 03:49:29.987158 1 request.go:481] Throttling request took 302.321948ms, request: PUT:https://10.100.0.1:443/api/v1/namespaces/kube-system/configmaps/cluster-autoscaler-status
I0813 03:49:30.784426 1 cloud_provider_builder.go:72] Building aws cloud provider.
I0813 03:49:30.784493 1 auto_scaling_groups.go:78] Registering ASG EKS-myeks-DefaultNodeGroup-NodeGroup-1CXC7XVC6QY2K
I0813 03:49:30.784499 1 auto_scaling_groups.go:139] Invalidating unowned instance cache
I0813 03:49:30.784506 1 auto_scaling_groups.go:153] Regenerating instance to ASG map for ASGs: [EKS-myeks-DefaultNodeGroup-NodeGroup-1CXC7XVC6QY2K]
I0813 03:49:31.493708 1 leaderelection.go:199] successfully renewed lease kube-system/cluster-autoscaler
I0813 03:49:31.794961 1 aws_manager.go:241] Refreshed ASG list, next refresh after 2018-08-13 03:50:31.79495307 +0000 UTC
I0813 03:49:31.795148 1 auto_scaling_groups.go:153] Regenerating instance to ASG map for ASGs: [EKS-myeks-DefaultNodeGroup-NodeGroup-1CXC7XVC6QY2K]
I0813 03:49:31.812340 1 main.go:228] Registered cleanup signal handler
I0813 03:49:33.515802 1 leaderelection.go:199] successfully renewed lease kube-system/cluster-autoscaler
I0813 03:49:35.525411 1 leaderelection.go:199] successfully renewed lease kube-system/cluster-autoscaler
I0813 03:49:37.537171 1 leaderelection.go:199] successfully renewed lease kube-system/cluster-autoscaler
I0813 03:49:39.583447 1 leaderelection.go:199] successfully renewed lease kube-system/cluster-autoscaler
I0813 03:49:41.690556 1 leaderelection.go:199] successfully renewed lease kube-system/cluster-autoscaler
You can manually increase the number of desired capacity (e.g. 8 in the example below).
And the cluster-autoscaler will probe the resource utilization and determine if the instances are unneeded and may be removed(scale in)
And overtime, the caluster-autoscaler will detect them and scale down the capacity and remove nodes being undeeded for 10 minutes
Let's create a deployment of Nginx with a replicaset of 1 Pod
$ kubectl apply -f templates/nginx-to-scaleout.yaml
deployment.extensions "nginx-to-scaleout" created
$ kubectl get deployment/nginx-to-scaleout
NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE
nginx-to-scaleout 1 1 1 1 8s
OK, let's scale out the replicaset to 20
$ kubectl get deployment/nginx-to-scaleout
NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE
nginx-to-scaleout 20 20 20 7 2m
Some pods will be in the Pending state, which triggers the cluster-autoscaler to scale out the EC2 fleet.
$ kubectl get po
NAME READY STATUS RESTARTS AGE
nginx-to-scaleout-7cb554c7d5-56vgm 0/1 Pending 0 39s
nginx-to-scaleout-7cb554c7d5-7bjrd 1/1 Running 0 39s
nginx-to-scaleout-7cb554c7d5-cb986 1/1 Running 0 2m
nginx-to-scaleout-7cb554c7d5-f7qrr 0/1 Pending 0 39s
nginx-to-scaleout-7cb554c7d5-fzjds 0/1 Pending 0 39s
nginx-to-scaleout-7cb554c7d5-g8ldm 0/1 Pending 0 39s
nginx-to-scaleout-7cb554c7d5-gf4x7 1/1 Running 0 39s
nginx-to-scaleout-7cb554c7d5-gszwf 0/1 Pending 0 39s
nginx-to-scaleout-7cb554c7d5-hzjzd 0/1 Pending 0 39s
nginx-to-scaleout-7cb554c7d5-jtfcp 1/1 Running 0 39s
nginx-to-scaleout-7cb554c7d5-nzdvl 0/1 Pending 0 39s
nginx-to-scaleout-7cb554c7d5-nzp9v 0/1 Pending 0 39s
nginx-to-scaleout-7cb554c7d5-pqqtt 1/1 Running 0 39s
nginx-to-scaleout-7cb554c7d5-q2hvb 1/1 Running 0 39s
nginx-to-scaleout-7cb554c7d5-q6lzp 1/1 Running 0 39s
nginx-to-scaleout-7cb554c7d5-sgxhg 1/1 Running 0 39s
nginx-to-scaleout-7cb554c7d5-szgb6 0/1 Pending 0 39s
nginx-to-scaleout-7cb554c7d5-tp2kl 0/1 Pending 0 39s
nginx-to-scaleout-7cb554c7d5-wvbdp 0/1 Pending 0 39s
nginx-to-scaleout-7cb554c7d5-xk26j 0/1 Pending 0 39s
You will see the autoscaling group starts scaling out immediately.
$ kubectl delete -f templates/nginx-to-scaleout.yaml
deployment.extensions "nginx-to-scaleout" deleted
$ kubelctl delete -f templates/eks-autoscaler.yaml
serviceaccount "cluster-autoscaler" deleted
clusterrole.rbac.authorization.k8s.io "cluster-autoscaler" deleted
role.rbac.authorization.k8s.io "cluster-autoscaler" deleted
clusterrolebinding.rbac.authorization.k8s.io "cluster-autoscaler" deleted
rolebinding.rbac.authorization.k8s.io "cluster-autoscaler" deleted
deployment.extensions "cluster-autoscaler" deleted
Cluster Autoscaler on AWS - https://github.com/kubernetes/autoscaler/blob/master/cluster-autoscaler/cloudprovider/aws/README.md
FAQ - https://github.com/kubernetes/autoscaler/blob/master/cluster-autoscaler/FAQ.md