Vertical Pod Autoscaler 0.2.0

NOTE: Notice on backwards compatibility

Highlights:

• VPA api changes:
  • Changed naming of some fields
  • New update mode added: Recreate (See Readme for details)
• State checkpointing for better handling of restart in recommender
• Scalability improvements in recommender
• Bug fixes

Cluster Autoscaler 1.3.0

Removes --configmap flag used for updating Cluster Autoscaler config via configmap.
Adds beta support for AKS.
Performance improvements in core Cluster Autoscaler. Performance improvements in GCP and AWS specific code.
Bugfixes.

Image: gcr.io/google_containers/cluster-autoscaler:v1.3.0

Cluster Autoscaler 1.3.0-beta.2

• Removed --configmap flag used for updating Cluster Autoscaler config via configmap.
• Performance improvements in core Cluster Autoscaler. Performance improvements in GCP and AWS specific code.
• Bugfixes.

Image: gcr.io/google_containers/cluster-autoscaler:v1.3.0-beta.2

Cluster Autoscaler 1.3.0-beta.1

Removes support for DynamicAutoscaler.

Image: gcr.io/google_containers/cluster-autoscaler:v1.3.0-beta.1

Cluster Autoscaler 1.2.2

Fixes draining nodes with StatefulSet pods.
Adds support for AKS cloud provider.

Image: gcr.io/google_containers/cluster-autoscaler:v1.2.2

Cluster Autoscaler 1.1.3

Fixes draining nodes with StatefulSet pods.
Fixes cleaning up old taints.

Image: gcr.io/google_containers/cluster-autoscaler:v1.1.3

Cluster Autoscaler 1.0.6

Fixes draining nodes with StatefulSet pods.
Fixes cleaning up old taints.

Image: gcr.io/google_containers/cluster-autoscaler:v1.0.6

Cluster Autoscaler 1.2.1

Minor fixes around kube-env parsing and persistent volume handling.

Image: gcr.io/google_containers/cluster-autoscaler:v1.2.1

Vertical Pod Autoscaler 0.1

First alpha release of Vertical Pod Autoscaler. Main features:

• Ability to work in a stand-alone mode, without persistent metrics storage.
• CRD-based API.
• Basic OutOfMemory error support.
• Experimental support for automatic pod request updating.

Cluster Autoscaler 1.2.0

Notable changes:

• Support Node Group autodiscovery on GCP (#462).
• Add Azure standard (VMAS) mode for kubernetes clusters deployed by acs-engine. More details here.
• More conservative and price-focused when scaling up node groups with GPUs.
• Fix handling of nodes failing to create due to overruning quota (#547).
• Use Debian base image.

Remarks:

• Topology Aware Volume Scheduling is turned off in 1.2.0. It should be enabled in 1.2.1.

Image: gcr.io/google_containers/cluster-autoscaler:v1.2.0