0.13

• #54: fix health endpoint to recover

Docker image is available:

docker run -it hjacobs/kube-aws-autoscaler:0.13 --help

0.12

• #45: add scale-down step functionality (fixed or percentage, thanks to @ichekrygin)

Docker image is available:

docker run -it hjacobs/kube-aws-autoscaler:0.12 --help

0.11

• #44: ignore failed pods that won't be restarted (thanks to @aermakov-zalando)

Docker image is available:

docker run -it hjacobs/kube-aws-autoscaler:0.11 --help

0.10

• #40: support more than 50 worker nodes (fix chunking to AWS API)
• #38: add --no-scale-down flag
• #35: add --enable-healthcheck-endpoint flag (expose /healthz on port 5000)

0.9

• #27: allow setting resource buffer values via environment variables

0.8

• #25: support clusters with more than 50 nodes (chunk input to DescribeAutoScalingInstances)

0.7

• #19: add timestamp to log output (important to make kubectl logs readable)
• #21: ignore master nodes by default (master ASG usually does not need any autoscaling)

0.6

• #17: add --buffer-spare-nodes and provision 1 extra spare node per ASG/AZ by default

0.5

• #11: use node Allocatable resources instead of raw capacity

0.4

• #6: scale up to MaxSize of ASG
• #10: ignore "ghost" pods (pods from terminated node)
• #9: more careful scale down: do not scale down on ASG activity or if some nodes are not ready yet

Docker image is available on Docker hub: https://hub.docker.com/r/hjacobs/kube-aws-autoscaler/