Releases: kubernetes/autoscaler
Cluster Autoscaler 1.13.0-alpha.1
- Add configurable delay for pod age before considering for scale-up via
--new-pod-scale-up-delay
param - Read AWS Region from EC2 Metadata
- Fetch ASG names in chunks of 50
- Ignore lo/fluentd-ds-ready when checking node similarity
- Support selecting which cloud providers are built in into CA binary
- Add Alibaba Cloud Provider
- Deprecate BasicEstimator
- Cloud provider SPI changes
- Add metrics in GCE cloud provider
- Add metrics in GKE cloud provider
- Various small bugfixes
- AWS load balancing improvements
- NodeGroupManager.CreateNodeGroup can create more than one node group.
gcr.io/google-containers/cluster-autoscaler:v1.13.0-alpha.1
Addon Resizer 1.8.4
- Fetch only metadata for nodes to optimize resource consumption.
Image: k8s.gcr.io/addon-resizer:1.8.4
Cluster Autoscaler 1.12.1
- Fix compatibility issues with k8s code changes.
gcr.io/google-containers/cluster-autoscaler:v1.12.1
Cluster Autoscaler 1.3.4
- Update list of AWS instance types
- Minor fixes
gcr.io/google-containers/cluster-autoscaler:v1.3.4
Cluster Autoscaler 1.3.3
- Fixes issue with node condition usage, which could lead to temporary provisioning of extra unneeded node.
gcr.io/google-containers/cluster-autoscaler:v1.3.3
Cluster Autoscaler 1.12.0
- Fix bug in scaleup related to ready condition handling
- Report reason why pod didn't trigger scale-up
- Fix fetching resource limits on GKE
- Fix GKE operation wait timeout
- Ignore resources with Cloud TPU prefix
- Update the aws ec2 instance types
- Adding the "cluster-autoscaler.kubernetes.io/safe-to-evict": "false"
- annotation to a pod prevents the cluster autoscaler from touching it.
- Add GPU-related scaled_up & scaled_down metrics
- Various fixes
Image: gcr.io/google-containers/cluster-autoscaler:v1.12.0
Cluster Autoscaler 1.2.3
Adds support for preventing pod eviction with safe-to-evict:false annotation.
Adds support for TPUs on GKE.
Fixes issues where AKS cluster may be unable to autoscale using old version of AKS API.
Improves allocatable memory node prediction on GKE.
Image: gcr.io/google-containers/cluster-autoscaler:v1.2.3
Cluster Autoscaler 1.12.0-beta.1
- Fix bug in scaleup related to ready condition handling
- Report reason why pod didn't trigger scale-up
- Fix fetching resource limits on GKE
- Fix GKE operation wait timeout
- Ignore resources with Cloud TPU prefix
- Update the aws ec2 instance types
- Adding the "cluster-autoscaler.kubernetes.io/safe-to-evict": "false"
annotation to a pod prevents the cluster autoscaler from touching it. - Add GPU-related scaled_up & scaled_down metrics
- Various fixes
Image: gcr.io/google-containers/cluster-autoscaler:v1.12.0-beta.1
Cluster Autoscaler 1.3.2
Fixes issues where AKS cluster may be unable to autoscale using old version of AKS API.
Improves allocatable memory node prediction on GKE.
Extends GKE TPU support to preemptibles.
Image: gcr.io/google-containers/cluster-autoscaler:v1.3.2
Cluster Autoscaler 1.2.3-beta.2
Adds support for preventing pod eviction with safe-to-evict:false annotation.
Adds support for TPUs on GKE.
Fixes issues where AKS cluster may be unable to autoscale using old version of AKS API.
Improves allocatable memory node prediction on GKE.
Image: gcr.io/google-containers/cluster-autoscaler:v1.2.3-beta.2