Does K8s support in-place autoscaling, which means update cpu resource on the fly. I want to change the cpu utilization of a pod without any termination(restarting or evcting).
I have look for this anwser for a long time. I see that the enhancement in K8s menting it is achieved in K8s, but when I asked this issue in slack of sig-autoscaling
, they told me it is not achieved now as in the PR, and it waits for other components(like kubelet or CRI) to supports auto-scaling on the fly.
Now I can’t tell whether K8s supports in-place autoscaling or not. Who can tell me the exact answer?
Thanks a lot!
Cluster information:
Kubernetes version:
Cloud being used: (put bare-metal if not on a public cloud)
Installation method:
Host OS: Ubuntu
CNI and version: any version
CRI and version: any version