provisioned nodeclaim getting into ready status is taking more time compared with older version

[dhanutalari]

Description

Observed Behavior:provisioned node getting into ready status is taking more time compared with older version

Expected Behavior:provisioned nodeclaim getting into ready status is need to be same time as older verion or less time.

Reproduction Steps (Please include YAML):

Versions:1.8.0

• Chart Version:1.0.1
• Kubernetes Version (kubectl version):1.3.2

• 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

[jmdeal]

There are a lot of factors that can effect this, most of which are outside of Karpenter's control:

• OS bootstrapping
• Networking configurations
• Unhealthy node components
• Transient instance launch failures
• etc

These are the steps Karpenter takes after a NodeClaim is created before marking it as ready:

• Launches instance (nodeclaim marked as launched upon success)
  • NodeClaim is deleted if there is a terminal failure (e.g. ICE)
  • For non-terminal errors the launch could be retried
• Instance registers with cluster
• All initialization conditions are met
  • Node is marked as ready by kubelet
  • All expected extended resources are registered
  • All startup taints are removed

To understand why there are delays, we would need to understand which step of this process the node / nodeclaim is getting stuck in.

/triage needs-information