Cluster Autoscaler 1.3.9

Changes:

• Fix AKS scaling issues for bulk deletion
• Update aws RefreshInterval to 1 minute
• Use debian-base-amd64:v1.0.0

Image: gcr.io/google-containers/cluster-autoscaler:v1.3.9

Cluster Autoscaler 1.14.1

Changes:

• Update aws RefreshInterval to 1 minute
• Use debian-base-amd64:v1.0.0

Image: gcr.io/google-containers/cluster-autoscaler:v1.14.1

Cluster Autoscaler 1.13.3

Changes:

• Fix AKS scaling issues for bulk deletion
• Update aws RefreshInterval to 1 minute
• Use debian-base-amd64:v1.0.0

Image: gcr.io/google-containers/cluster-autoscaler:v1.13.3

Cluster Autoscaler 1.12.4

Changes:

• Fix AKS scaling issues for bulk deletion
• Update aws RefreshInterval to 1 minute
• Use debian-base-amd64:v1.0.0

Image: gcr.io/google-containers/cluster-autoscaler:v1.12.4

Vertical Pod Autoscaler 0.4.1

Notable changes:

• Correct garbage collection of containers with no samples.
• Validation if targetRef is a top level controller.
• Monitoring metrics for unsupported VPA configurations.
• VPA condition indicating if any pods are matched.
• Enhanced logging.

Vertical Pod Autoscaler 0.5.0

API related changes

• [IMPORTANT] autoscaling.k8s.io/v1beta1 API is no longer supported in this release
  • existing v1beta1 VerticalPodAutoscaler objects will get an empty recommendation,
  • validation introduced to disallow creating objects without targetRef defined,
  • please migrate your objects to v1beta2 prior to upgrading.

Other changes

• Configuration options for Prometheus setup added.
• Configuration options for Admission Controller registration.

Cluster Autoscaler 1.14.0

Changes since 1.13.0:

• Update k8s/kubernetes dependency to 1.14
• Update base debian (0.4.1)
• (AWS) Update AWS EC2 instance type catalog
• (AWS) Report ASGs using unknown AWS EC2 instance types
• Add flags to ignore daemonsets and mirror pods when calculating resource utilization of a node
• (GCE) Handle possible stockout/quota scale-up errors, by failing scalup
• (Azure) Add cache for resource IDs of vmss instances
• Tainting unneeded nodes as PreferNoSchedule
• Removing nominatedNodeName annotation and moving to pod.Status.NominatedNodeName
• Added a new method that uses pod packing to filter schedulable pods
• (Azure) Fix windows name parsing for Azure VMAS nodes
• Account for kernel reserved memory in capacity calculations
• Cache exemplar ready node for each node group
• (AWS) Adding ability to override allocatable resources via ASG tags.
• Avoid panic in random expander
• (GCE) Added migTargetSizeCache in GceCache to cache TargetSize calls for each GCEMig for each loop of the CA. The cache is invalidated during refresh that occurs during the start of each loop and during MIG size change (SetMigSize and DeleteInstances)
• Fixing minor error handling bug in static autoscaler
• (Azure) Fix AKS scaling issues for bulk deletion
• (AWS) Allow custom AWS region overrides
• Other smaller bug-fixes/cleanups

Image: gcr.io/google-containers/cluster-autoscaler:v1.14.0

Cluster Autoscaler 1.14.0-beta.2

TODO: Changelog
Since cluster-autoscaler:v1.14.0-beta.1:

• Fix anti-affinity predicates checking (proper scheduler simulator initialization)
• Allow custom AWS region overrides
• Fix AKS scaling issues for bulk deletion
• Soft tainting fix
• fix max cluster size calculation on scale up

Image: gcr.io/google-containers/cluster-autoscaler:v1.14.0-beta.2

Cluster Autoscaler 1.14.0-beta.1

TODO: Changelog

Image: gcr.io/google-containers/cluster-autoscaler:v1.14.0-beta.1

Cluster Autoscaler 1.3.8

• Update base debian image to newest version
• Fix bug related to instance name case sensitivity for Azure (#1608)
• Fix windows name parsing for Azure VMAS nodes

Image: gcr.io/google-containers/cluster-autoscaler:v1.3.8