Kubernetes Cluster Autoscaler Support (Experimental)

By default, Longhorn blocks Kubernetes Cluster Autoscaler from scaling down nodes because:

  • Longhorn creates PodDisruptionBudgets for all engine and replica instance-manager pods.
  • Longhorn instance manager pods have strict PodDisruptionBudgets.
  • Longhorn instance manager pods are not backed by a Kubernetes built-in workload controller .
  • Longhorn pods are using local storage volume mounts.

For more information, see What types of pods can prevent CA from removing a node?

If you want to unblock the Kubernetes Cluster Autoscaler scaling, you can set the setting Kubernetes Cluster Autoscaler Enabled.

When this setting is enabled, Longhorn will retain the least instance-manager PodDisruptionBudget as possible. Each volume will have at least one replica under the protection of an instance-manager PodDisruptionBudget while no redundant PodDisruptionBudget blocking the Cluster Autoscaler from from scaling down.

When this setting is enabled, Longhorn will also add cluster-autoscaler.kubernetes.io/safe-to-evict annotation to Longhorn workloads that are not backed by a Kubernetes built-in workload controller or are using local storage mounts.

Warning: Replica rebuilding could be expensive because nodes with reusable replicas could get removed by the Kubernetes Cluster Autoscaler.


© 2019-2022 Longhorn Authors | Documentation Distributed under CC-BY-4.0


© 2022 The Linux Foundation. All rights reserved. The Linux Foundation has registered trademarks and uses trademarks. For a list of trademarks of The Linux Foundation, please see our Trademark Usage page.