v1.6 to v1.7

Follow the Regular Upgrading Process.

Upgrading Notable Changes

API Changes

  • Introduced more printcolumn for FederatedHPA, including reference, minpods, maxpods and replicas.
  • Introduced new API CronFederatedHPA to scale the workloads in specific time.
  • Introduced Preemption to both PropagationPolicy and ClusterPropagationPolicy to declare the behaviors of preemption.
  • Introduced ConflictResolution to both PropagationPolicy and ClusterPropagationPolicy to declare how potential conflict should be handled.
  • Introduced a new field Zones for Cluster to represent multiple zones of a member cluster, the old filed zone is deprecated.

karmadactl

  • Introduced --wait-component-ready-timeout flag to specify the component installation timeout.
  • Introduced top command.

karmada-controller-manager

  • You are advised to enableclusterrole-aggregation controller to grant ClusterRole/admin with Karmada resource permission.
  • Introduced a new feature-gate --feature-gates=PropagationPolicyPreemption=true to enable policy preemption by priority.
  • Introduced --cluster-cache-sync-timeout flag to specify the sync timeout of the control plane cache in addition to the member cluster’s cache.
  • Introduced a LabelSelector field to DependentObjectReference.

karmada-scheduler

  • Introduced new scheduling condition reasons: NoClusterFit, SchedulerError, Unschedulable, Success.

karmada-metrics-adapter

  • Introduced karmada-metrics-adapter to addons to utilize FederatedHPA scaling workloads across multiple clusters, it can be installed by karmadactl and karmada-operator.

Please refer to v1.7.0 Release Notes for more details.