CCE Cluster Autoscaler

Introduction

Autoscaler is an important Kubernetes controller. It supports microservice scaling and is key to serverless design.

When the CPU or memory usage of a microservice is too high, horizontal pod autoscaling is triggered to add pods to reduce the load. These pods can be automatically reduced when the load is low, allowing the microservice to run as efficiently as possible.

CCE simplifies the creation, upgrade, and manual scaling of Kubernetes clusters, in which traffic loads change over time. To balance resource usage and workload performance of nodes, Kubernetes introduces the Autoscaler add-on to automatically adjust the number of nodes a cluster based on the resource usage required for workloads deployed in the cluster. For details, see Creating a Node Scaling Policy.

Open source community: https://github.com/kubernetes/autoscaler

How the Add-on Works

Autoscaler controls auto scale-out and scale-in.

• Auto scale-out
  You can choose either of the following methods:

  • If pods in a cluster cannot be scheduled due to insufficient worker nodes, cluster scaling is triggered to add nodes. The nodes to be added have the same specification as configured for the node pool to which the nodes belong.
    Auto scale-out will be performed when:

    • Node resources are insufficient.
    • No node affinity policy is set in the pod scheduling configuration. If a node has been configured as an affinity node for pods, no node will not be automatically added when pods cannot be scheduled. For details about how to configure the node affinity policy, see Scheduling Policies (Affinity/Anti-affinity).
  • When the cluster meets the node scaling policy, cluster scale-out is also triggered. For details, see Creating a Node Scaling Policy.

  

  The add-on follows the "No Less, No More" policy. For example, if three cores are required for creating a pod and the system supports four-core and eight-core nodes, Autoscaler will preferentially create a four-core node.

• Auto scale-in
  When a cluster node is idle for a period of time (10 minutes by default), cluster scale-in is triggered, and the node is automatically deleted. However, a node cannot be deleted from a cluster if the following pods exist:

  • Pods that do not meet specific requirements set in Pod Disruption Budgets (PodDisruptionBudget)
  • Pods that cannot be scheduled to other nodes due to constraints such as affinity and anti-affinity policies
  • Pods that have the cluster-autoscaler.kubernetes.io/safe-to-evict: 'false' annotation
  • Pods (except those created by DaemonSets in the kube-system namespace) that exist in the kube-system namespace on the node
  • Pods that are not created by the controller (Deployment/ReplicaSet/job/StatefulSet)

  

  When a node meets the scale-in conditions, Autoscaler adds the DeletionCandidateOfClusterAutoscaler taint to the node in advance to prevent pods from being scheduled to the node. After the Autoscaler add-on is uninstalled, if the taint still exists on the node, manually delete it.

Notes and Constraints

• Ensure that there are sufficient resources for installing the add-on.
• The default node pool does not support auto scaling. For details, see Description of DefaultPool.
• Node scale-in will cause PVC/PV data loss for the local PVs associated with the node. These PVCs and PVs cannot be restored or used again. In a node scale-in, a pod that uses the local PV will be evicted from the node. A new pod will be created, but it remains in a pending state because the label of the PVC bound to it conflicts with the node label.
• When Autoscaler is used, some taints or annotations may affect auto scaling. Therefore, do not use the following taints or annotations in clusters:
  • ignore-taint.cluster-autoscaler.kubernetes.io: The taint works on nodes. Kubernetes-native Autoscaler supports protection against abnormal scale outs and periodically evaluates the proportion of available nodes in the cluster. When the proportion of non-ready nodes exceeds 45%, protection will be triggered. In this case, all nodes with the ignore-taint.cluster-autoscaler.kubernetes.io taint in the cluster are filtered out from the Autoscaler template and recorded as non-ready nodes, which affect cluster scaling.
  • cluster-autoscaler.kubernetes.io/enable-ds-eviction: The annotation works on pods, which determines whether DaemonSet pods can be evicted by Autoscaler. For details, see Well-Known Labels, Annotations and Taints.

Installing the Add-on

1. Log in to the CCE console and click the cluster name to access the cluster console. Choose Add-ons in the navigation pane, locate CCE Cluster Autoscaler on the right, and click Install.
2. On the Install Add-on page, configure the specifications.

   Table 1 Specifications configuration

   Parameter	Description
   Pods	Number of pods for the add-on. High availability is not possible with a single pod. If an error occurs on the node where the add-on instance runs, the add-on will fail.
   Containers	Adjust the number of the Autoscaler pods and their CPU and memory quotas based on the cluster scale. For details, see Table 2.

   Table 2 Recommended Autoscaler quotas

   Nodes	Pods	Requested vCPUs	vCPU Limit	Requested Memory	Memory Limit
   50	2	1000m	1000m	1000 MiB	1000 MiB
   200	2	4000m	4000m	2000 MiB	2000 MiB
   1000	2	8000m	8000m	8000 MiB	8000 MiB
   2000	2	8000m	8000m	8000 MiB	8000 MiB

3. Configure the add-on parameters.

   Table 3 Parameters

   Parameter	Description
   Total Nodes	Maximum number of nodes that can be managed by the cluster, within which cluster scale-out is performed.
   Total CPUs	Maximum sum of CPU cores of all nodes in a cluster, within which cluster scale-out is performed.
   Total Memory (GiB)	Maximum sum of memory of all nodes in a cluster, within which cluster scale-out is performed.

4. Configure scheduling policies for the add-on.
   

   • Scheduling policies do not take effect on add-on instances of the DaemonSet type.
   • When configuring multi-AZ deployment or node affinity, ensure that there are nodes meeting the scheduling policy and that resources are sufficient in the cluster. Otherwise, the add-on cannot run.

   Table 4 Configurations for add-on scheduling

   Parameter	Description
   Multi AZ	Preferred: Deployment pods of the add-on will be preferentially scheduled to nodes in different AZs. If all the nodes in the cluster are deployed in the same AZ, the pods will be scheduled to that AZ. Equivalent mode: Deployment pods of the add-on are evenly scheduled to the nodes in the cluster in each AZ. If a new AZ is added, you are advised to increase add-on pods for cross-AZ HA deployment. With the Equivalent multi-AZ deployment, the difference between the number of add-on pods in different AZs will be less than or equal to 1. If resources in one of the AZs are insufficient, pods cannot be scheduled to that AZ. Required: Deployment pods of the add-on will be forcibly scheduled to nodes in different AZs. If there are fewer AZs than pods, the extra pods will fail to run.
   Node Affinity	Not configured: Node affinity is disabled for the add-on. Node Affinity: Specify the nodes where the add-on is deployed. If you do not specify the nodes, the add-on will be randomly scheduled based on the default cluster scheduling policy. Specified Node Pool Scheduling: Specify the node pool where the add-on is deployed. If you do not specify the node pool, the add-on will be randomly scheduled based on the default cluster scheduling policy. Custom Policies: Enter the labels of the nodes where the add-on is to be deployed for more flexible scheduling policies. If you do not specify node labels, the add-on will be randomly scheduled based on the default cluster scheduling policy. If multiple custom affinity policies are configured, ensure that there are nodes that meet all the affinity policies in the cluster. Otherwise, the add-on cannot run.
   Toleration	Using both taints and tolerations allows (not forcibly) the add-on Deployment to be scheduled to a node with the matching taints, and controls the Deployment eviction policies after the node where the Deployment is located is tainted. The add-on adds the default tolerance policy for the node.kubernetes.io/not-ready and node.kubernetes.io/unreachable taints, respectively. The tolerance time window is 60s. For details, see Configuring Tolerance Policies.

5. After the configuration is complete, click Install.

Components

Scale-In Cool-Down Period

Scale-in cooling intervals can be configured in the node pool settings and the Autoscaler add-on settings.

Scale-in cooling interval configured in a node pool

This interval indicates the period during which nodes added to the current node pool after a scale-out cannot be deleted. This setting takes effect in the entire node pool.

Scale-in cooling interval configured in the Autoscaler add-on

The interval after a scale-out indicates the period during which the entire cluster cannot be scaled in after the Autoscaler add-on triggers a scale-out (due to the unschedulable pods, metrics, and scaling policies). This setting takes effect in the entire cluster.

The interval after a node is deleted indicates the period during which the cluster cannot be scaled in after the auto scaling add-on triggers a scale-in. This setting takes effect in the entire cluster.

The interval after a failed scale-in indicates the period during which the cluster cannot be scaled in after the Autoscaler add-on triggers scale-in. This interval takes effect at the cluster level.