Cluster Autoscaler 1.14.8
Changes since 1.14.7
General:
- Fix potential SEGV in updateReadinessStat
AWS:
- Change the providerID regex to support Fargate
- Add instructions on how to tag resources for the scale from 0 case
- Support arbitrary custom resource building template
Azure:
- Fix VMSS scale down issues when updating target sizes
- Fix cloud provider config parsing
- Add missing prefix in user-agent
- Fix panic when VMSS doesn't exist
- Add VM cache
- Don't return an error for nodes without providerId
- Decrease number of calls to VMSS API
- Make VMSS info cache TTL configurable
- Fix a bug in VM cache when multiple agentpool share the same cache
- Change scale-up failure handling to properly handle node group backoff
- Ensure only capacity is changed when making a PUT on VMSS
Images:
- eu.gcr.io/k8s-artifacts-prod/autoscaling/cluster-autoscaler:v1.14.8
- us.gcr.io/k8s-artifacts-prod/autoscaling/cluster-autoscaler:v1.14.8
- asia.gcr.io/k8s-artifacts-prod/autoscaling/cluster-autoscaler:v1.14.8