Delay in Pod Scheduling After Scale-Up Action

[Idan-Lazar]

Description

Observed Behavior:
Pods are scheduled approximately 30 seconds after a scale-up action.

Expected Behavior:
I want the pods to be scheduled immediately after the scale-up action.

Reproduction Steps (Please include YAML):

Versions:

• Chart Version: karpenter-1.0.0
• Kubernetes Version (kubectl version):
  Client Version: v1.28.2
  Server Version: v1.29.7-eks-a18cd3a

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

[jigisha620]

How quickly are you expecting the pods to get scheduled? Pod would start running only when the capacity is provisioned for it to run. And that could take time, but that's not something controlled by Karpenter.

[Idan-Lazar]

Before Karpenter, it takes less than a second.
For example, let's assume that I have a deployment with two pods of Ngnix, when a pod is terminated a new pod will show after 30 seconds. It's a long time, and without Karpenter it doesn't take long.