Alauda Container Platform
简体中文

容器逃逸防护策略

本指南演示如何配置 Kyverno，通过阻止可能导致容器突破隔离边界的高风险容器配置，来防止容器逃逸攻击。

目录

什么是容器逃逸防护？

容器逃逸防护涉及检测并阻止可能允许攻击者突破容器隔离、访问宿主机系统的危险容器配置。包括：

  • 特权容器：以提升权限运行的容器
  • 宿主机命名空间访问：容器共享宿主机的 PID、网络或 IPC 命名空间
  • 宿主机路径挂载：容器挂载宿主机文件系统路径
  • 危险的能力：容器拥有过多的 Linux 能力
  • 宿主机端口访问：容器绑定宿主机网络端口

快速开始

1. 阻止特权容器

apiVersion: kyverno.io/v1
kind: ClusterPolicy
metadata:
  name: disallow-privileged-containers
  annotations:
    policies.kyverno.io/title: Disallow Privileged Containers
    policies.kyverno.io/category: Pod Security Standards (Baseline)
    policies.kyverno.io/severity: medium
    policies.kyverno.io/subject: Pod
    policies.kyverno.io/description: >-
      Privileged mode disables most security mechanisms and must not be allowed.
spec:
  validationFailureAction: Enforce
  background: true
  rules:
    - name: privileged-containers
      match:
        any:
        - resources:
            kinds:
            - Pod
      validate:
        message: >-
          Privileged mode is disallowed. The fields spec.containers[*].securityContext.privileged,
          spec.initContainers[*].securityContext.privileged, and spec.ephemeralContainers[*].securityContext.privileged 
          must be unset or set to false.
        pattern:
          spec:
            =(ephemeralContainers):
              - =(securityContext):
                  =(privileged): "false"
            =(initContainers):
              - =(securityContext):
                  =(privileged): "false"
            containers:
              - =(securityContext):
                  =(privileged): "false"

2. 测试策略

# 应用策略
kubectl apply -f disallow-privileged-containers.yaml

# 尝试创建特权容器（应失败）
cat <<EOF | kubectl apply -f -
apiVersion: v1
kind: Pod
metadata:
  name: test-privileged
spec:
  containers:
  - name: nginx
    image: nginx
    securityContext:
      privileged: true
EOF

# 尝试创建普通容器（应成功）
kubectl run test-normal --image=nginx

# 清理资源
kubectl delete pod test-privileged test-normal --ignore-not-found

核心容器逃逸防护策略

策略 1：禁止访问宿主机命名空间

防止容器访问宿主机命名空间：

apiVersion: kyverno.io/v1
kind: ClusterPolicy
metadata:
  name: disallow-host-namespaces
  annotations:
    policies.kyverno.io/title: Disallow Host Namespaces
    policies.kyverno.io/category: Pod Security Standards (Baseline)
    policies.kyverno.io/severity: medium
    policies.kyverno.io/subject: Pod
    policies.kyverno.io/description: >-
      Host namespaces (Process ID namespace, Inter-Process Communication namespace, and 
      network namespace) allow access to shared information and can be used to elevate 
      privileges. Pods should not be allowed access to host namespaces.
spec:
  validationFailureAction: Enforce
  background: true
  rules:
    - name: host-namespaces
      match:
        any:
        - resources:
            kinds:
            - Pod
      validate:
        message: >-
          Sharing the host namespaces is disallowed. The fields spec.hostNetwork,
          spec.hostIPC, and spec.hostPID must be unset or set to false.
        pattern:
          spec:
            =(hostPID): "false"
            =(hostIPC): "false"
            =(hostNetwork): "false"

策略 2：禁止宿主机路径挂载

阻止容器挂载宿主机文件系统路径：

apiVersion: kyverno.io/v1
kind: ClusterPolicy
metadata:
  name: disallow-host-path
  annotations:
    policies.kyverno.io/title: Disallow Host Path
    policies.kyverno.io/category: Pod Security Standards (Baseline)
    policies.kyverno.io/severity: medium
    policies.kyverno.io/subject: Pod,Volume
    policies.kyverno.io/description: >-
      HostPath volumes let Pods use host directories and volumes in containers.
      Using host resources can be used to access shared data or escalate privileges
      and should not be allowed.
spec:
  validationFailureAction: Enforce
  background: true
  rules:
    - name: host-path
      match:
        any:
        - resources:
            kinds:
            - Pod
      validate:
        message: >-
          HostPath volumes are forbidden. The field spec.volumes[*].hostPath must be unset.
        pattern:
          spec:
            =(volumes):
              - X(hostPath): "null"

策略 3：禁止宿主机端口

防止容器绑定宿主机网络端口：

apiVersion: kyverno.io/v1
kind: ClusterPolicy
metadata:
  name: disallow-host-ports
  annotations:
    policies.kyverno.io/title: Disallow Host Ports
    policies.kyverno.io/category: Pod Security Standards (Baseline)
    policies.kyverno.io/severity: medium
    policies.kyverno.io/subject: Pod
    policies.kyverno.io/description: >-
      Access to host ports allows potential snooping of network traffic and should not be
      allowed, or at minimum restricted to a known list.
spec:
  validationFailureAction: Enforce
  background: true
  rules:
    - name: host-ports-none
      match:
        any:
        - resources:
            kinds:
            - Pod
      validate:
        message: >-
          Use of host ports is disallowed. The fields spec.containers[*].ports[*].hostPort,
          spec.initContainers[*].ports[*].hostPort, and spec.ephemeralContainers[*].ports[*].hostPort
          must either be unset or set to 0.
        pattern:
          spec:
            =(ephemeralContainers):
              - =(ports):
                  - =(hostPort): 0
            =(initContainers):
              - =(ports):
                  - =(hostPort): 0
            containers:
              - =(ports):
                  - =(hostPort): 0

策略 4：禁止危险能力

阻止容器添加危险的 Linux 能力：

apiVersion: kyverno.io/v1
kind: ClusterPolicy
metadata:
  name: disallow-capabilities-strict
  annotations:
    policies.kyverno.io/title: Disallow Capabilities (Strict)
    policies.kyverno.io/category: Pod Security Standards (Restricted)
    policies.kyverno.io/severity: medium
    policies.kyverno.io/subject: Pod
    policies.kyverno.io/description: >-
      Adding capabilities other than `NET_BIND_SERVICE` is disallowed. In addition,
      all containers must explicitly drop `ALL` capabilities.
spec:
  validationFailureAction: Enforce
  background: true
  rules:
    - name: require-drop-all
      match:
        any:
        - resources:
            kinds:
            - Pod
      preconditions:
        all:
        - key: "{{ request.operation || 'BACKGROUND' }}"
          operator: NotEquals
          value: DELETE
      validate:
        message: >-
          Containers must drop `ALL` capabilities.
        foreach:
        - list: request.object.spec.[ephemeralContainers, initContainers, containers][]
          deny:
            conditions:
              all:
              - key: ALL
                operator: AnyNotIn
                value: "{{ element.securityContext.capabilities.drop || `[]` }}"
    - name: adding-capabilities
      match:
        any:
        - resources:
            kinds:
            - Pod
      preconditions:
        all:
        - key: "{{ request.operation || 'BACKGROUND' }}"
          operator: NotEquals
          value: DELETE
      validate:
        message: >-
          Any capabilities added other than NET_BIND_SERVICE are disallowed.
        foreach:
        - list: request.object.spec.[ephemeralContainers, initContainers, containers][]
          deny:
            conditions:
              any:
              - key: "{{ element.securityContext.capabilities.add || `[]` }}"
                operator: AnyNotIn
                value:
                - NET_BIND_SERVICE

策略 5：要求非 root 用户运行容器

确保容器以非 root 用户身份运行：

apiVersion: kyverno.io/v1
kind: ClusterPolicy
metadata:
  name: require-run-as-nonroot
  annotations:
    policies.kyverno.io/title: Require Run As Non-Root User
    policies.kyverno.io/category: Pod Security Standards (Restricted)
    policies.kyverno.io/severity: medium
    policies.kyverno.io/subject: Pod
    policies.kyverno.io/description: >-
      Containers must run as a non-root user. This policy ensures runAsNonRoot is set to true.
spec:
  validationFailureAction: Enforce
  background: true
  rules:
    - name: run-as-non-root
      match:
        any:
        - resources:
            kinds:
            - Pod
      validate:
        message: >-
          Running as root is not allowed. Either the field spec.securityContext.runAsNonRoot 
          must be set to true, or the field spec.containers[*].securityContext.runAsNonRoot 
          must be set to true.
        anyPattern:
        - spec:
            securityContext:
              runAsNonRoot: "true"
        - spec:
            containers:
            - securityContext:
                runAsNonRoot: "true"

高级场景

场景 1：环境特定策略

针对不同环境设置不同的安全级别：

apiVersion: kyverno.io/v1
kind: ClusterPolicy
metadata:
  name: environment-container-security
spec:
  validationFailureAction: Enforce
  background: true
  rules:
    # 生产环境：严格安全
    - name: production-strict-security
      match:
        any:
        - resources:
            kinds:
            - Pod
            namespaces:
            - production
            - prod-*
      validate:
        message: "Production environments require strict container security"
        pattern:
          spec:
            =(hostPID): "false"
            =(hostIPC): "false"
            =(hostNetwork): "false"
            securityContext:
              runAsNonRoot: "true"
            containers:
            - securityContext:
                privileged: "false"
                runAsNonRoot: "true"
                capabilities:
                  drop:
                  - ALL
    
    # 开发环境：更宽松但仍安全
    - name: development-basic-security
      match:
        any:
        - resources:
            kinds:
            - Pod
            namespaces:
            - development
            - dev-*
            - staging
      validate:
        message: "Development environments require basic container security"
        pattern:
          spec:
            =(hostPID): "false"
            =(hostIPC): "false"
            containers:
            - securityContext:
                =(privileged): "false"

场景 2：工作负载特定例外

允许特定工作负载的受控例外：

apiVersion: kyverno.io/v1
kind: ClusterPolicy
metadata:
  name: workload-specific-security
spec:
  validationFailureAction: Enforce
  background: true
  rules:
    - name: system-workloads-exception
      match:
        any:
        - resources:
            kinds:
            - Pod
      exclude:
        any:
        - resources:
            namespaces:
            - kube-system
            - kyverno
        - resources:
            kinds:
            - Pod
            names:
            - "monitoring-*"
            - "logging-*"
      validate:
        message: "Container security policies apply to application workloads"
        pattern:
          spec:
            =(hostNetwork): "false"
            containers:
            - securityContext:
                =(privileged): "false"

测试与验证

测试特权容器

# 此操作应被阻止
cat <<EOF | kubectl apply -f -
apiVersion: v1
kind: Pod
metadata:
  name: test-privileged
spec:
  containers:
  - name: test
    image: nginx
    securityContext:
      privileged: true
EOF

测试宿主机命名空间访问

# 此操作应被阻止
cat <<EOF | kubectl apply -f -
apiVersion: v1
kind: Pod
metadata:
  name: test-host-network
spec:
  hostNetwork: true
  containers:
  - name: test
    image: nginx
EOF

测试宿主机路径挂载

# 此操作应被阻止
cat <<EOF | kubectl apply -f -
apiVersion: v1
kind: Pod
metadata:
  name: test-hostpath
spec:
  containers:
  - name: test
    image: nginx
    volumeMounts:
    - name: host-vol
      mountPath: /host
  volumes:
  - name: host-vol
    hostPath:
      path: /
EOF

测试安全合规容器

# 此操作应被允许
cat <<EOF | kubectl apply -f -
apiVersion: v1
kind: Pod
metadata:
  name: test-secure
spec:
  securityContext:
    runAsNonRoot: true
    runAsUser: 1000
  containers:
  - name: test
    image: nginx
    securityContext:
      allowPrivilegeEscalation: false
      capabilities:
        drop:
        - ALL
      readOnlyRootFilesystem: true
      runAsNonRoot: true
      runAsUser: 1000
EOF

最佳实践

1. 从审计模式开始

spec:
  validationFailureAction: Audit  # 先以警告形式提示，不阻止

2. 排除系统命名空间

exclude:
  any:
  - resources:
      namespaces:
      - kube-system
      - kyverno
      - kube-public

以 PDF 格式查看完整文档