Kubernetes VPA 垂直 Pod 自动伸缩完整指南

# 检查 Kubernetes 版本
kubectl version --short

# 检查集群节点状态
kubectl get nodes

# 检查是否已安装 Metrics Server
kubectl get deployment metrics-server -n kube-system

# 下载 OpenSSL 1.1.1k 源码
wget http://mirrors.ibiblio.org/openssl/source/openssl-1.1.1k.tar.gz

# 解压并进入目录
tar xf openssl-1.1.1k.tar.gz
cd openssl-1.1.1k/

# 配置编译选项
./config --prefix=/usr/local/openssl --openssldir=/usr/local/openssl

# 编译安装（使用所有 CPU 核心加速编译）
make -j$(nproc)
sudo make install

# 备份原有 OpenSSL
sudo mv /usr/bin/openssl{,.bak}

# 创建软链接
sudo ln -s /usr/local/openssl/bin/openssl /usr/bin/openssl
sudo ln -s /usr/local/openssl/lib/libssl.so.1.1 /usr/lib64/libssl.so.1.1
sudo ln -s /usr/local/openssl/lib/libcrypto.so.1.1 /usr/lib64/libcrypto.so.1.1

# 更新动态链接库缓存
sudo ldconfig

# 验证安装
openssl version

# 下载并安装 Metrics Server v0.5.0
kubectl apply -f https://github.com/kubernetes-sigs/metrics-server/releases/download/v0.5.0/components.yaml

# 编辑 Metrics Server 部署配置
kubectl edit deployment metrics-server -n kube-system

spec:
  template:
    spec:
      containers:
      - name: metrics-server
        args:
        - --cert-dir=/tmp
        - --secure-port=443
        - --kubelet-preferred-address-types=InternalIP,ExternalIP,Hostname
        - --kubelet-use-node-status-port
        - --metric-resolution=15s
        - --kubelet-insecure-tls  # 添加此行，跳过 TLS 验证

# 等待 Pod 启动完成
watch kubectl get pods -l k8s-app=metrics-server -n kube-system

# 验证 Metrics Server 功能
kubectl top nodes

# 检查 Pod 资源使用情况
kubectl top pods --all-namespaces

# 克隆 Kubernetes Autoscaler 项目
git clone https://github.com/kubernetes/autoscaler.git

# 进入 VPA 目录
cd autoscaler/vertical-pod-autoscaler/

# 查看可用版本（可选）
git tag | grep vpa | tail -5

# 执行安装脚本（默认安装 v0.9.2）
./hack/vpa-up.sh

# 检查 VPA 相关 Pod 状态
kubectl get pods -n kube-system | grep vpa

# 验证 VPA 自定义资源定义
kubectl get crd | grep verticalpodautoscaler

# 查看 VPA CRD 详情
kubectl describe crd verticalpodautoscalers.autoscaling.k8s.io

# 查看 VPA 相关服务
kubectl get svc -n kube-system | grep vpa

# 查看 VPA Webhook 配置
kubectl get validatingwebhookconfiguration | grep vpa
kubectl get mutatingwebhookconfiguration | grep vpa

# 创建专用命名空间
kubectl create namespace vpa-demo

# 应用示例配置
kubectl apply -f examples/hamster.yaml -n vpa-demo

# hamster.yaml 内容解析
apiVersion: apps/v1
kind: Deployment
metadata:
  name: hamster
spec:
  replicas: 2
  selector:
    matchLabels:
      app: hamster
  template:
    metadata:
      labels:
        app: hamster
    spec:
      containers:
      - name: hamster
        image: k8s.gcr.io/ubuntu-slim:0.1
        resources:
          requests:
            cpu: 100m      # 初始 CPU 请求：100 毫核
            memory: 50Mi   # 初始内存请求：50MB
        command: ["/bin/sh"]
        args:
          - "-c"
          - "while true; do timeout 10s yes >/dev/null; sleep 5s; done"
---
apiVersion: autoscaling.k8s.io/v1
kind: VerticalPodAutoscaler
metadata:
  name: hamster-vpa
spec:
  targetRef:
    apiVersion: "apps/v1"
    kind: Deployment
    name: hamster
  updatePolicy:
    updateMode: "Auto"    # 自动更新模式

# 获取 Pod 名称
POD_NAME=$(kubectl get pods -n vpa-demo -l app=hamster -o jsonpath='{.items[0].metadata.name}')

# 查看初始资源配置
kubectl get pod $POD_NAME -n vpa-demo -o yaml | grep -A 5 'resources:'

resources:
  requests:
    cpu: 100m
    memory: 50Mi

# 查看 VPA 状态和推荐
kubectl describe vpa hamster-vpa -n vpa-demo

# 监控 Pod 变化
watch kubectl get pods -n vpa-demo

# 获取新 Pod 名称
NEW_POD_NAME=$(kubectl get pods -n vpa-demo -l app=hamster -o jsonpath='{.items[0].metadata.name}')

# 查看更新后的资源配置
kubectl get pod $NEW_POD_NAME -n vpa-demo -o yaml | grep -A 5 'resources:'

resources:
  requests:
    cpu: 548m        # CPU 请求增加到 548 毫核
    memory: 262144k  # 内存请求增加到约 256MB

cat <<EOF | kubectl apply -f -
apiVersion: autoscaling.k8s.io/v1
kind: VerticalPodAutoscaler
metadata:
  name: hamster-vpa-off
  namespace: vpa-demo
spec:
  targetRef:
    apiVersion: "apps/v1"
    kind: Deployment
    name: hamster
  updatePolicy:
    updateMode: "Off"  # 仅提供推荐，不自动更新
EOF

cat <<EOF | kubectl apply -f -
apiVersion: autoscaling.k8s.io/v1
kind: VerticalPodAutoscaler
metadata:
  name: hamster-vpa-initial
  namespace: vpa-demo
spec:
  targetRef:
    apiVersion: "apps/v1"
    kind: Deployment
    name: hamster
  updatePolicy:
    updateMode: "Initial"  # 仅对新创建的 Pod 生效
EOF

cat <<EOF | kubectl apply -f -
apiVersion: autoscaling.k8s.io/v1
kind: VerticalPodAutoscaler
metadata:
  name: hamster-vpa-policy
  namespace: vpa-demo
spec:
  targetRef:
    apiVersion: "apps/v1"
    kind: Deployment
    name: hamster
  updatePolicy:
    updateMode: "Auto"
  resourcePolicy:
    containerPolicies:
    - containerName: hamster
      minAllowed:
        cpu: 100m
        memory: 50Mi
      maxAllowed:
        cpu: 1000m
        memory: 500Mi
      controlledResources: ["cpu", "memory"]
EOF

# 查看详细的 VPA 状态
kubectl describe vpa hamster-vpa -n vpa-demo

# 查看 VPA 推荐的 JSON 格式
kubectl get vpa hamster-vpa -n vpa-demo -o json | jq '.status.recommendation'

# 查看 Recommender 日志
kubectl logs -n kube-system -l app=vpa-recommender

# 查看 Updater 日志
kubectl logs -n kube-system -l app=vpa-updater

# 查看 Admission Controller 日志
kubectl logs -n kube-system -l app=vpa-admission-controller

# 查看 VPA 推荐指标
curl http://<vpa-recommender-ip>:8080/metrics

# 查看 VPA 更新指标
curl http://<vpa-updater-ip>:8080/metrics

resourcePolicy:
  containerPolicies:
  - containerName: app
    minAllowed:
      cpu: 100m      # 防止资源过低
      memory: 128Mi
    maxAllowed:
      cpu: 2000m     # 防止资源过高
      memory: 4Gi
    controlledResources: ["cpu", "memory"]
    controlledValues: "RequestsAndLimits"  # 同时控制 requests 和 limits

# 推荐配置 Prometheus 监控规则
groups:
- name: vpa.rules
  rules:
  - alert: VPARecommendationChanged
    expr: increase(vpa_recommender_recommendations_total[5m]) > 0
    labels:
      severity: info
    annotations:
      summary: "VPA recommendation changed for {{ $labels.vpa_name }}"

# 检查 VPA 状态
kubectl get vpa --all-namespaces

# 查看 VPA 事件
kubectl get events --field-selector involvedObject.kind=VerticalPodAutoscaler

# 检查 Webhook 配置
kubectl get validatingwebhookconfiguration vpa-webhook-config -o yaml

# 测试 VPA API
kubectl api-resources | grep verticalpodautoscaler