5주차 - EKS Autoscaling - Karpenter

Karpenter

 

CA의 경우 AWS ASG와 CA 모두 동작을 해야 하기 때문에 시간이 오래 걸림

ASG의 경우 주체가 AWS이고, CA의 경우 EKS가 주체이기 때문에 각 각 확인을 해야 함

 

Karpenter의 경우 ASG를 사용하지 않고 Karpenter 자체적으로 동작하기 때문에 CA보다 훨씬 빠르게 동작

 

 

 

Karpenter 설치

• 기존 배포한 EKS 삭제 후 새로운 EKS 배포

# 변수 정보 확인
export | egrep 'ACCOUNT|AWS_' | egrep -v 'SECRET|KEY'

# 변수 설정
export KARPENTER_NAMESPACE="kube-system"
export K8S_VERSION="1.29"
export KARPENTER_VERSION="0.35.2"
export TEMPOUT=$(mktemp)
export ARM_AMI_ID="$(aws ssm get-parameter --name /aws/service/eks/optimized-ami/${K8S_VERSION}/amazon-linux-2-arm64/recommended/image_id --query Parameter.Value --output text)"
export AMD_AMI_ID="$(aws ssm get-parameter --name /aws/service/eks/optimized-ami/${K8S_VERSION}/amazon-linux-2/recommended/image_id --query Parameter.Value --output text)"
export GPU_AMI_ID="$(aws ssm get-parameter --name /aws/service/eks/optimized-ami/${K8S_VERSION}/amazon-linux-2-gpu/recommended/image_id --query Parameter.Value --output text)"
export AWS_PARTITION="aws"
export CLUSTER_NAME="${USER}-karpenter-demo"
echo "export CLUSTER_NAME=$CLUSTER_NAME" >> /etc/profile
echo $KARPENTER_VERSION $CLUSTER_NAME $AWS_DEFAULT_REGION $AWS_ACCOUNT_ID $TEMPOUT $ARM_AMI_ID $AMD_AMI_ID $GPU_AMI_ID

# CloudFormation 스택으로 IAM Policy, Role(KarpenterNodeRole-myeks2) 생성 : 3분 정도 소요
curl -fsSL https://raw.githubusercontent.com/aws/karpenter-provider-aws/v"${KARPENTER_VERSION}"/website/content/en/preview/getting-started/getting-started-with-karpenter/cloudformation.yaml  > "${TEMPOUT}" \
&& aws cloudformation deploy \
  --stack-name "Karpenter-${CLUSTER_NAME}" \
  --template-file "${TEMPOUT}" \
  --capabilities CAPABILITY_NAMED_IAM \
  --parameter-overrides "ClusterName=${CLUSTER_NAME}"

# 클러스터 생성 : myeks2 EKS 클러스터 생성 19분 정도 소요
eksctl create cluster -f - <<EOF
---
apiVersion: eksctl.io/v1alpha5
kind: ClusterConfig
metadata:
  name: ${CLUSTER_NAME}
  region: ${AWS_DEFAULT_REGION}
  version: "${K8S_VERSION}"
  tags:
    karpenter.sh/discovery: ${CLUSTER_NAME}

iam:
  withOIDC: true
  serviceAccounts:
  - metadata:
      name: karpenter
      namespace: "${KARPENTER_NAMESPACE}"
    roleName: ${CLUSTER_NAME}-karpenter
    attachPolicyARNs:
    - arn:${AWS_PARTITION}:iam::${AWS_ACCOUNT_ID}:policy/KarpenterControllerPolicy-${CLUSTER_NAME}
    roleOnly: true

iamIdentityMappings:
- arn: "arn:${AWS_PARTITION}:iam::${AWS_ACCOUNT_ID}:role/KarpenterNodeRole-${CLUSTER_NAME}"
  username: system:node:{{EC2PrivateDNSName}}
  groups:
  - system:bootstrappers
  - system:nodes

managedNodeGroups:
- instanceType: m5.large
  amiFamily: AmazonLinux2
  name: ${CLUSTER_NAME}-ng
  desiredCapacity: 2
  minSize: 1
  maxSize: 10
  iam:
    withAddonPolicies:
      externalDNS: true
EOF

# eks 배포 확인
eksctl get cluster
eksctl get nodegroup --cluster $CLUSTER_NAME
eksctl get iamidentitymapping --cluster $CLUSTER_NAME
eksctl get iamserviceaccount --cluster $CLUSTER_NAME
eksctl get addon --cluster $CLUSTER_NAME

• Karpenter 설치

# karpenter 설치
helm install karpenter oci://public.ecr.aws/karpenter/karpenter --version "${KARPENTER_VERSION}" --namespace "${KARPENTER_NAMESPACE}" --create-namespace \
  --set "serviceAccount.annotations.eks\.amazonaws\.com/role-arn=${KARPENTER_IAM_ROLE_ARN}" \
  --set "settings.clusterName=${CLUSTER_NAME}" \
  --set "settings.interruptionQueue=${CLUSTER_NAME}" \
  --set controller.resources.requests.cpu=1 \
  --set controller.resources.requests.memory=1Gi \
  --set controller.resources.limits.cpu=1 \
  --set controller.resources.limits.memory=1Gi \
  --wait
 
# 확인
kubectl get-all -n $KARPENTER_NAMESPACE
kubectl get all -n $KARPENTER_NAMESPACE
kubectl get crd | grep karpenter

# APi 변경
v1alpha5/Provisioner → v1beta1/NodePool
v1alpha1/AWSNodeTemplate → v1beta1/EC2NodeClass 
v1alpha5/Machine → v1beta1/NodeClaim

 

 

 

실습시나리오 / Node 생성 및 삭제

 

1. NodePool 생성

2. Test Pod 배포

3. Replicas 증가에 따른 Node 생성 확인

4. Node 삭제 확인

 

1. NodePool 생성

• NodePool을 통해 어떤 Instance를 생성 할 지 지정
• EC2NodeClass를 통해 Role 및 Subnet, SG 설정

cat <<EOF | envsubst | kubectl apply -f -
apiVersion: karpenter.sh/v1beta1
kind: NodePool
metadata:
  name: default
spec:
  template:
    spec:
      requirements:
        - key: kubernetes.io/arch
          operator: In
          values: ["amd64"]
        - key: kubernetes.io/os
          operator: In
          values: ["linux"]
        - key: karpenter.sh/capacity-type
          operator: In
          values: ["spot"]
        - key: karpenter.k8s.aws/instance-category
          operator: In
          values: ["c", "m", "r"]
        - key: karpenter.k8s.aws/instance-generation
          operator: Gt
          values: ["2"]
      nodeClassRef:
        apiVersion: karpenter.k8s.aws/v1beta1
        kind: EC2NodeClass
        name: default
  limits:
    cpu: 1000
  disruption:
    consolidationPolicy: WhenUnderutilized
    expireAfter: 720h # 30 * 24h = 720h
---
apiVersion: karpenter.k8s.aws/v1beta1
kind: EC2NodeClass
metadata:
  name: default
spec:
  amiFamily: AL2 # Amazon Linux 2
  role: "KarpenterNodeRole-${CLUSTER_NAME}" # replace with your cluster name
  subnetSelectorTerms:
    - tags:
        karpenter.sh/discovery: "${CLUSTER_NAME}" # replace with your cluster name
  securityGroupSelectorTerms:
    - tags:
        karpenter.sh/discovery: "${CLUSTER_NAME}" # replace with your cluster name
  amiSelectorTerms:
    - id: "${ARM_AMI_ID}"
    - id: "${AMD_AMI_ID}"
#   - id: "${GPU_AMI_ID}" # <- GPU Optimized AMD AMI 
#   - name: "amazon-eks-node-${K8S_VERSION}-*" # <- automatically upgrade when a new AL2 EKS Optimized AMI is released. This is unsafe for production workloads. Validate AMIs in lower environments before deploying them to production.
EOF

# 확인
kubectl get nodepool,ec2nodeclass

2. Test Pod 배포

# pause 파드 1개에 CPU 1개 최소 보장 할당
cat <<EOF | kubectl apply -f -
apiVersion: apps/v1
kind: Deployment
metadata:
  name: inflate
spec:
  replicas: 0
  selector:
    matchLabels:
      app: inflate
  template:
    metadata:
      labels:
        app: inflate
    spec:
      terminationGracePeriodSeconds: 0
      containers:
        - name: inflate
          image: public.ecr.aws/eks-distro/kubernetes/pause:3.7
          resources:
            requests:
              cpu: 1
EOF

 

3. Replicas 증가에 따른 Node 생성 확인

• replicas 수를 5로 변경

kubectl scale deployment inflate --replicas 5
kubectl logs -f -n "${KARPENTER_NAMESPACE}" -l app.kubernetes.io/name=karpenter -c controller
kubectl logs -f -n "${KARPENTER_NAMESPACE}" -l app.kubernetes.io/name=karpenter -c controller | jq '.'

 

• Pod Pending 확인

 

3-1. Node 추가 확인 - Spot Instance

 

4. Node 삭제 확인

• deployment 삭제로 인하여 Node 삭제 확인

# Now, delete the deployment. After a short amount of time, Karpenter should terminate the empty nodes due to consolidation.
kubectl delete deployment inflate && date
kubectl logs -f -n "${KARPENTER_NAMESPACE}" -l app.kubernetes.io/name=karpenter -c controller

 

 

 

실습시나리오 / Disruption(Consolidation) 확인

Node 최적화로 Karpenter가 자원이 남는 Node가 있으면 삭제하거나 또는 스펙을 줄임 

 

1. 기존 NodePool 삭제 및 신규 NodePool 생성 

2. Test Pod 배포

3. Node 추가 확인 ( Spot Instance)

4. Node 스펙 Down 확인 (Spot-to-Spot Consolidation)

 

 

1. 기존 NodePool 삭제 및 신규 NodePool 생성 

• Expiration 만료 : 기본 720시간(30일) 후 인스턴스를 자동으로 만료하여 강제로 노드를 최신 상태로 유지
• Drift 드리프트 : 구성 변경 사항(NodePool, EC2NodeClass)를 감지하여 필요한 변경 사항을 적용
• Consolidation 통합 : 비용 효율적인 컴퓨팅 최적화

# 기존 nodepool 삭제
kubectl delete nodepool,ec2nodeclass default

# v0.34.0 부터 featureGates 에 spotToSpotConsolidation 활성화로 사용 가능
helm upgrade karpenter -n kube-system oci://public.ecr.aws/karpenter/karpenter --reuse-values --set settings.featureGates.spotToSpotConsolidation=true

# Create a Karpenter NodePool and EC2NodeClass
cat <<EOF > nodepool.yaml
apiVersion: karpenter.sh/v1beta1
kind: NodePool
metadata:
  name: default
spec:
  template:
    metadata:
      labels:
        intent: apps
    spec:
      nodeClassRef:
        name: default
      requirements:
        - key: karpenter.sh/capacity-type
          operator: In
          values: ["spot"]
        - key: karpenter.k8s.aws/instance-category
          operator: In
          values: ["c","m","r"]
        - key: karpenter.k8s.aws/instance-size
          operator: NotIn
          values: ["nano","micro","small","medium"]
        - key: karpenter.k8s.aws/instance-hypervisor
          operator: In
          values: ["nitro"]
  limits:
    cpu: 100
    memory: 100Gi
  disruption:
    consolidationPolicy: WhenUnderutilized
---
apiVersion: karpenter.k8s.aws/v1beta1
kind: EC2NodeClass
metadata:
  name: default
spec:
  amiFamily: Bottlerocket
  subnetSelectorTerms:          
    - tags:
        karpenter.sh/discovery: "${CLUSTER_NAME}" # replace with your cluster name
  securityGroupSelectorTerms:
    - tags:
        karpenter.sh/discovery: "${CLUSTER_NAME}" # replace with your cluster name
  role: "KarpenterNodeRole-${CLUSTER_NAME}" # replace with your cluster name
  tags:
    Name: karpenter.sh/nodepool/default
    IntentLabel: "apps"
EOF
kubectl apply -f nodepool.yaml

 

2. Test Pod 배포

# Deploy a sample workload
cat <<EOF > inflate.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: inflate
spec:
  replicas: 5
  selector:
    matchLabels:
      app: inflate
  template:
    metadata:
      labels:
        app: inflate
    spec:
      nodeSelector:
        intent: apps
      containers:
        - name: inflate
          image: public.ecr.aws/eks-distro/kubernetes/pause:3.2
          resources:
            requests:
              cpu: 1
              memory: 1.5Gi
EOF
kubectl apply -f inflate.yaml

 

3. Node 추가 확인 ( Spot Instance)

 

• 기존 replicas 수가 5개로 Pod Pending에 따른 Spot instance 추가

 

4. Node 스펙 Down 확인 (Spot-to-Spot Consolidation)

• 스팟 간 통합이 이루어지려면 Karpenter에는 최소 15개의 인스턴스 유형이 포함된 다양한 인스턴스 구성(연습에 정의된 NodePool 예제 참조)
  • helm upgrade karpenter -n kube-system oci://public.ecr.aws/karpenter/karpenter --reuse-values --set settings.featureGates.spotToSpotConsolidation=true

# Scale in a sample workload to observe consolidation
# To invoke a Karpenter consolidation event scale, inflate the deployment to 1. Run the following command:
kubectl scale --replicas=1 deployment/inflate

 

kubectl -n kube-system logs -l app.kubernetes.io/name=karpenter --all-containers=true -f --tail=20
kubectl get nodes -L karpenter.sh/nodepool -L node.kubernetes.io/instance-type -L topology.kubernetes.io/zone -L karpenter.sh/capacity-type
kubectl get node --label-columns=eks.amazonaws.com/capacityType,karpenter.sh/capacity-type
kubectl get node --label-columns=node.kubernetes.io/instance-type,topology.kubernetes.io/zone

 

• Spot Instance의 Spec Down 확인
  • c6gd.2xlarge → c6gd.large