Karpenter Schedule for Node Disruption

[tarunmittal24]

Hi I am facing a issue in which the node got refreshed, one's the new ami is being released by the aws team, as i have not mentioned the ami-selector terms this is a expected behaviour, but as per security i need to update image every 30 days, that i have handled with the expireAfter: 720h in the nodepool config, But the case is, if i mentioned the ami-selector terms and put the specific image-ami-id, even after it got expired in the 30 days it will take the same image.

Any way to control this, like want the new nodes will comes after x days, with the new ami as that particular point of time and want that it should happened during the specific schedule like on weekends [Saturday (2-3 am)], not on weekdays.

My Current Config

Node-pool.yaml

apiVersion: karpenter.sh/v1beta1
kind: NodePool
metadata:
  name: core
spec:
  disruption:
    consolidateAfter: Never
    expireAfter: 720h
    consolidationPolicy: WhenEmpty
    budgets:
      - nodes: 10%
      - schedule: "0 0 * * mon-fri"
        duration: 24h
        nodes: "0"
  limits:
    cpu: "12"
    memory: 45Gi
  template:
    metadata:
      labels:
        node: core
    spec:
      taints:
        - key: "node"
          value: "core"
          effect: "NoSchedule"
      nodeClassRef:
        apiVersion: karpenter.k8s.aws/v1beta1
        kind: EC2NodeClass
        name: core
      kubelet: 
        maxPods: 110
      requirements:
      - key: node.kubernetes.io/instance-type
        operator: In
        values: 
          - m5a.large
          - m6a.large
          - m6i.large
          - m5.large
          - m4.large
      - key: kubernetes.io/os
        operator: In
        values:
        - linux
      - key: kubernetes.io/arch
        operator: In
        values:
        - amd64
      - key: karpenter.sh/capacity-type	
        operator: In	
        values:
          - on-demand

node-class.yaml

apiVersion: karpenter.k8s.aws/v1beta1
kind: EC2NodeClass
metadata:
 name: core
spec:
 amiFamily: AL2
 metadataOptions:
   httpEndpoint: enabled
   httpTokens: required
 subnetSelectorTerms:
   - tags:
       karpenter.sh/zfh-prod/team: "infra"
       aws-cdk:subnet-name: "EKSPrivate"
 securityGroupSelectorTerms:
   - id: "sg-009b2e138f9263f3d" 
 role: "PROD-EKS-KarpenterNode"  
 tags:
   Name: PROD-EKS/karpenter.sh/core
 metadataOptions:
   httpEndpoint: enabled
   httpProtocolIPv6: disabled
   httpPutResponseHopLimit: 2
   httpTokens: required

[k8s-ci-robot]

This issue is currently awaiting triage.

If Karpenter contributors determines this is a relevant issue, they will accept it by applying the triage/accepted label and provide further guidance.

The triage/accepted label can be added by org members by writing /triage accepted in a comment.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[njtran]

What I think you want is Node Disruption Budgets, which was released in v0.34. https://karpenter.sh/docs/concepts/disruption/#disruption-budgets.

Going to close this, feel free to re-open if you think your feature request is different.