1. Documentation
×

Applying autoscaling to an OKD cluster

Applying autoscaling to an OKD cluster involves deploying a cluster autoscaler and then deploying machine autoscalers for each machine type in your cluster.

You can configure the cluster autoscaler only in clusters where the machine API is operational.

About the cluster autoscaler

The cluster autoscaler adjusts the size of an OKD cluster to meet its current deployment needs. It uses declarative, Kubernetes-style arguments to provide infrastructure management that does not rely on objects of a specific cloud provider. The cluster autoscaler has a cluster scope, and is not associated with a particular namespace.

The cluster autoscaler increases the size of the cluster when there are pods that fail to schedule on any of the current worker nodes due to insufficient resources or when another node is necessary to meet deployment needs. The cluster autoscaler does not increase the cluster resources beyond the limits that you specify.

The cluster autoscaler computes the total memory, CPU, and GPU on all nodes the cluster, even though it does not manage the control plane nodes. These values are not single-machine oriented. They are an aggregation of all the resources in the entire cluster. For example, if you set the maximum memory resource limit, the cluster autoscaler includes all the nodes in the cluster when calculating the current memory usage. That calculation is then used to determine if the cluster autoscaler has the capacity to add more worker resources.

Ensure that the maxNodesTotal value in the ClusterAutoscaler resource definition that you create is large enough to account for the total possible number of machines in your cluster. This value must encompass the number of control plane machines and the possible number of compute machines that you might scale to.

Every 10 seconds, the cluster autoscaler checks which nodes are unnecessary in the cluster and removes them. The cluster autoscaler considers a node for removal if the following conditions apply:

  • The node utilization is less than the node utilization level threshold for the cluster. The node utilization level is the sum of the requested resources divided by the allocated resources for the node. If you do not specify a value in the ClusterAutoscaler custom resource, the cluster autoscaler uses a default value of 0.5, which corresponds to 50% utilization.

  • The cluster autoscaler can move all pods running on the node to the other nodes. The Kubernetes scheduler is responsible for scheduling pods on the nodes.

  • The cluster autoscaler does not have scale down disabled annotation.

If the following types of pods are present on a node, the cluster autoscaler will not remove the node:

  • Pods with restrictive pod disruption budgets (PDBs).

  • Kube-system pods that do not run on the node by default.

  • Kube-system pods that do not have a PDB or have a PDB that is too restrictive.

  • Pods that are not backed by a controller object such as a deployment, replica set, or stateful set.

  • Pods with local storage.

  • Pods that cannot be moved elsewhere because of a lack of resources, incompatible node selectors or affinity, matching anti-affinity, and so on.

  • Unless they also have a "cluster-autoscaler.kubernetes.io/safe-to-evict": "true" annotation, pods that have a "cluster-autoscaler.kubernetes.io/safe-to-evict": "false" annotation.

For example, you set the maximum CPU limit to 64 cores and configure the cluster autoscaler to only create machines that have 8 cores each. If your cluster starts with 30 cores, the cluster autoscaler can add up to 4 more nodes with 32 cores, for a total of 62.

If you configure the cluster autoscaler, additional usage restrictions apply:

  • Do not modify the nodes that are in autoscaled node groups directly. All nodes within the same node group have the same capacity and labels and run the same system pods.

  • Specify requests for your pods.

  • If you have to prevent pods from being deleted too quickly, configure appropriate PDBs.

  • Confirm that your cloud provider quota is large enough to support the maximum node pools that you configure.

  • Do not run additional node group autoscalers, especially the ones offered by your cloud provider.

The horizontal pod autoscaler (HPA) and the cluster autoscaler modify cluster resources in different ways. The HPA changes the deployment’s or replica set’s number of replicas based on the current CPU load. If the load increases, the HPA creates new replicas, regardless of the amount of resources available to the cluster. If there are not enough resources, the cluster autoscaler adds resources so that the HPA-created pods can run. If the load decreases, the HPA stops some replicas. If this action causes some nodes to be underutilized or completely empty, the cluster autoscaler deletes the unnecessary nodes.

The cluster autoscaler takes pod priorities into account. The Pod Priority and Preemption feature enables scheduling pods based on priorities if the cluster does not have enough resources, but the cluster autoscaler ensures that the cluster has resources to run all pods. To honor the intention of both features, the cluster autoscaler includes a priority cutoff function. You can use this cutoff to schedule "best-effort" pods, which do not cause the cluster autoscaler to increase resources but instead run only when spare resources are available.

Pods with priority lower than the cutoff value do not cause the cluster to scale up or prevent the cluster from scaling down. No new nodes are added to run the pods, and nodes running these pods might be deleted to free resources.

Cluster autoscaling is supported for the platforms that have machine API available on it.

Configuring the cluster autoscaler

First, deploy the cluster autoscaler to manage automatic resource scaling in your OKD cluster.

Because the cluster autoscaler is scoped to the entire cluster, you can make only one cluster autoscaler for the cluster.

ClusterAutoscaler resource definition

This ClusterAutoscaler resource definition shows the parameters and sample values for the cluster autoscaler.

apiVersion: "autoscaling.openshift.io/v1"
kind: "ClusterAutoscaler"
metadata:
  name: "default"
spec:
  podPriorityThreshold: -10 (1)
  resourceLimits:
    maxNodesTotal: 24 (2)
    cores:
      min: 8 (3)
      max: 128 (4)
    memory:
      min: 4 (5)
      max: 256 (6)
    gpus:
      - type: nvidia.com/gpu (7)
        min: 0 (8)
        max: 16 (9)
      - type: amd.com/gpu
        min: 0
        max: 4
  scaleDown: (10)
    enabled: true (11)
    delayAfterAdd: 10m (12)
    delayAfterDelete: 5m (13)
    delayAfterFailure: 30s (14)
    unneededTime: 5m (15)
    utilizationThreshold: "0.4" (16)
1 Specify the priority that a pod must exceed to cause the cluster autoscaler to deploy additional nodes. Enter a 32-bit integer value. The podPriorityThreshold value is compared to the value of the PriorityClass that you assign to each pod.
2 Specify the maximum number of nodes to deploy. This value is the total number of machines that are deployed in your cluster, not just the ones that the autoscaler controls. Ensure that this value is large enough to account for all of your control plane and compute machines and the total number of replicas that you specify in your MachineAutoscaler resources.
3 Specify the minimum number of cores to deploy in the cluster.
4 Specify the maximum number of cores to deploy in the cluster.
5 Specify the minimum amount of memory, in GiB, in the cluster.
6 Specify the maximum amount of memory, in GiB, in the cluster.
7 Optional: Specify the type of GPU node to deploy. Only nvidia.com/gpu and amd.com/gpu are valid types.
8 Specify the minimum number of GPUs to deploy in the cluster.
9 Specify the maximum number of GPUs to deploy in the cluster.
10 In this section, you can specify the period to wait for each action by using any valid ParseDuration interval, including ns, us, ms, s, m, and h.
11 Specify whether the cluster autoscaler can remove unnecessary nodes.
12 Optional: Specify the period to wait before deleting a node after a node has recently been added. If you do not specify a value, the default value of 10m is used.
13 Optional: Specify the period to wait before deleting a node after a node has recently been deleted. If you do not specify a value, the default value of 0s is used.
14 Optional: Specify the period to wait before deleting a node after a scale down failure occurred. If you do not specify a value, the default value of 3m is used.
15 Optional: Specify the period before an unnecessary node is eligible for deletion. If you do not specify a value, the default value of 10m is used.
16 Optional: Specify the node utilization level below which an unnecessary node is eligible for deletion. The node utilization level is the sum of the requested resources divided by the allocated resources for the node, and must be a value greater than "0" but less than "1". If you do not specify a value, the cluster autoscaler uses a default value of "0.5", which corresponds to 50% utilization. This value must be expressed as a string.

When performing a scaling operation, the cluster autoscaler remains within the ranges set in the ClusterAutoscaler resource definition, such as the minimum and maximum number of cores to deploy or the amount of memory in the cluster. However, the cluster autoscaler does not correct the current values in your cluster to be within those ranges.

The minimum and maximum CPUs, memory, and GPU values are determined by calculating those resources on all nodes in the cluster, even if the cluster autoscaler does not manage the nodes. For example, the control plane nodes are considered in the total memory in the cluster, even though the cluster autoscaler does not manage the control plane nodes.

Deploying the cluster autoscaler

To deploy the cluster autoscaler, you create an instance of the ClusterAutoscaler resource.

Procedure

  1. Create a YAML file for the ClusterAutoscaler resource that contains the customized resource definition.

  2. Create the resource in the cluster:

    $ oc create -f <filename>.yaml (1)
    1 <filename> is the name of the resource file that you customized.

Next steps

  • After you configure the cluster autoscaler, you must configure at least one machine autoscaler.

About the machine autoscaler

The machine autoscaler adjusts the number of Machines in the machine sets that you deploy in an OKD cluster. You can scale both the default worker machine set and any other machine sets that you create. The machine autoscaler makes more Machines when the cluster runs out of resources to support more deployments. Any changes to the values in MachineAutoscaler resources, such as the minimum or maximum number of instances, are immediately applied to the machine set they target.

You must deploy a machine autoscaler for the cluster autoscaler to scale your machines. The cluster autoscaler uses the annotations on machine sets that the machine autoscaler sets to determine the resources that it can scale. If you define a cluster autoscaler without also defining machine autoscalers, the cluster autoscaler will never scale your cluster.

Configuring the machine autoscalers

After you deploy the cluster autoscaler, deploy MachineAutoscaler resources that reference the machine sets that are used to scale the cluster.

You must deploy at least one MachineAutoscaler resource after you deploy the ClusterAutoscaler resource.

You must configure separate resources for each machine set. Remember that machine sets are different in each region, so consider whether you want to enable machine scaling in multiple regions. The machine set that you scale must have at least one machine in it.

MachineAutoscaler resource definition

This MachineAutoscaler resource definition shows the parameters and sample values for the machine autoscaler.

apiVersion: "autoscaling.openshift.io/v1beta1"
kind: "MachineAutoscaler"
metadata:
  name: "worker-us-east-1a" (1)
  namespace: "openshift-machine-api"
spec:
  minReplicas: 1 (2)
  maxReplicas: 12 (3)
  scaleTargetRef: (4)
    apiVersion: machine.openshift.io/v1beta1
    kind: MachineSet (5)
    name: worker-us-east-1a (6)
1 Specify the machine autoscaler name. To make it easier to identify which machine set this machine autoscaler scales, specify or include the name of the machine set to scale. The machine set name takes the following form: <clusterid>-<machineset>-<region>.
2 Specify the minimum number machines of the specified type that must remain in the specified zone after the cluster autoscaler initiates cluster scaling. If running in AWS, GCP, Azure, OpenStack, or vSphere, this value can be set to 0. For other providers, do not set this value to 0.

You can save on costs by setting this value to 0 for use cases such as running expensive or limited-usage hardware that is used for specialized workloads, or by scaling a machine set with extra large machines. The cluster autoscaler scales the machine set down to zero if the machines are not in use.

Do not set the spec.minReplicas value to 0 for the three compute machine sets that are created during the OKD installation process for an installer provisioned infrastructure.
3 Specify the maximum number machines of the specified type that the cluster autoscaler can deploy in the specified zone after it initiates cluster scaling. Ensure that the maxNodesTotal value in the ClusterAutoscaler resource definition is large enough to allow the machine autoscaler to deploy this number of machines.
4 In this section, provide values that describe the existing machine set to scale.
5 The kind parameter value is always MachineSet.
6 The name value must match the name of an existing machine set, as shown in the metadata.name parameter value.

Deploying the machine autoscaler

To deploy the machine autoscaler, you create an instance of the MachineAutoscaler resource.

Procedure

  1. Create a YAML file for the MachineAutoscaler resource that contains the customized resource definition.

  2. Create the resource in the cluster:

    $ oc create -f <filename>.yaml (1)
    1 <filename> is the name of the resource file that you customized.

Additional resources