如何了解 ClusterResourcePlacement 和 ResourcePlacement 的状态

2025-08-13

使用 Azure Kubernetes Fleet Manager 时，了解（CRP）资源的状态 ClusterResourcePlacement 对于监视部署进度和排查问题至关重要。本文提供了一个全面的指南，用于解释 Fleet Manager 针对群集范围和命名空间范围的放置报告的状态字段和条件。

先决条件

你有一个具有中心群集和一个或多个成员群集的机群管理器。如果没有，请参阅创建 Azure Kubernetes Fleet Manager 资源并加入成员群集。
有权访问 Fleet Manager 中心群集。有关详细信息，请参阅访问 Azure Kubernetes Fleet Manager 中心群集的 Kubernetes API。
你已部署了至少一个 ClusterResourcePlacement API 对象，用于将资源放入群中。如果没有，请参阅 “使用群集资源放置”跨多个群集部署工作负荷。

放置状态结构的概述

对象 ClusterResourcePlacement 不仅包含有关放置的描述性规范，还包含放置作的状态。状态部分提供有关以下内容的详细信息：

通过条件表示的总体放置状态
位置选择的资源
按群集放置状态通过条件表示
Failed、 drifted和 diffed 每个群集中的位置

若要查看放置状态，请使用以下命令：

kubectl describe clusterresourceplacement <placement-name>

或者获取原始 YAML 输出：

kubectl get clusterresourceplacement <placement-name> -o yaml

顶级状态字段

状态部分包含以下顶级字段：

selectedResources：位置选择的资源列表
条件：总体放置条件数组
observedResourceIndex：当前资源快照的索引
placementStatuses：每群集放置状态信息

以下部分详细介绍了每个字段。

了解所选资源

该 selectedResources 字段列出放置选择的所有资源。此字段允许检查放置中是否包含预期的资源。下面是一个示例：

selectedResources:
- kind: Namespace
  name: test
  version: v1
- kind: ConfigMap
  name: test-config
  namespace: test
  version: v1
  envelope:
    name: example-envelope
    namespace: test
    type: ResourceEnvelope
- kind: Deployment
  name: web-app
  namespace: test
  group: apps
  version: v1
- kind: Service
  name: web-service
  namespace: test
  version: v1
- kind: Secret
  name: app-secrets
  namespace: test
  version: v1

每个资源条目包括：

组：API 组（核心资源的空）
版本：API 版本（例如，v1） v1beta1
kind：资源类型
名称：资源名称
命名空间：命名空间（用于命名空间资源）
信封：如果资源包装在信封中，则还会提供信封元数据，其中包括：
- 名称：信封的名称
- 命名空间：信封的命名空间
- 类型：信封的类型（例如 ResourceEnvelope）

了解放置条件

该 conditions 数组提供有关整个放置的高级状态信息。每个条件都遵循 Kubernetes 通用定义，该定义具有以下标准字段：

类型：条件类型（以下部分所述）
状态： True、 False或 Unknown
原因：条件的简短原因代码
消息：人工可读说明
lastTransitionTime：条件上次更改时
observedGeneration：设置条件时放置的生成

ClusterResourcePlacement 条件类型

以下条件类型可用于 ClusterResourcePlacement：

ClusterResourcePlacementScheduled

指示放置是否成功计划到目标群集。

True：根据放置策略选择所有必需的群集
False：计划失败（例如，群集可用不足）
未知：计划决策处于挂起状态

conditions:
- type: ClusterResourcePlacementScheduled
  status: "True"
  reason: SchedulingPolicyFulfilled
  message: "found all the clusters needed as specified by the scheduling policy"
  lastTransitionTime: "2023-11-10T08:14:52Z"
  observedGeneration: 5

ClusterResourcePlacementRolloutStarted

指示是否在所选群集中启动推出。

True：资源已开始向计划群集推出
False：尚未启动推出
未知：推出决策处于挂起状态

conditions:
- type: ClusterResourcePlacementRolloutStarted
  status: "True"
  reason: RolloutStarted
  message: "All 3 cluster(s) start rolling out the latest resource"
  lastTransitionTime: "2023-11-10T08:15:30Z"
  observedGeneration: 5

ClusterResourcePlacementOverridden

指示是否成功应用资源替代。

True：处理所有适用的替代
False：某些替代无法应用
未知：替代处理处于挂起状态

conditions:
- type: ClusterResourcePlacementOverridden
  status: "True"
  reason: NoOverrideSpecified
  message: "No override rules are configured for the selected resources"
  lastTransitionTime: "2023-11-10T08:15:45Z"
  observedGeneration: 5

ClusterResourcePlacementWorkSynchronized

指示是否在中心群集的每个群集命名空间中创建工作对象。

True：所有工作对象都同步
False：工作同步失败或不完整
未知：工作同步挂起

conditions:
- type: ClusterResourcePlacementWorkSynchronized  
  status: "True"
  reason: SynchronizeSucceeded
  message: "All 2 cluster(s) are synchronized to the latest resources on the hub cluster"
  lastTransitionTime: "2023-11-10T08:23:43Z"
  observedGeneration: 5

ClusterResourcePlacementApplied

指示是否将所有资源成功应用于成员群集。

True：已成功应用于所有目标群集的所有资源
False：某些资源无法应用（检查 failedPlacements）
未知：应用作挂起

conditions:
- type: ClusterResourcePlacementApplied
  status: "True"
  reason: ApplySucceeded
  message: "The selected resources are successfully applied to 3 clusters"
  lastTransitionTime: "2023-11-10T08:16:15Z"
  observedGeneration: 5

ClusterResourcePlacementAvailable

指示放置的资源是否全部可用，并且已准备好在成员群集上。

True：所有目标群集上都提供所有资源
False：某些资源尚不可用
未知：可用性检查处于挂起状态

conditions:
- type: ClusterResourcePlacementAvailable
  status: "True"
  reason: ResourceAvailable
  message: "The selected resources in 3 clusters are available now"
  lastTransitionTime: "2023-11-10T08:16:30Z"
  observedGeneration: 5

ClusterResourcePlacementDiffReported

指示是否报告配置差异（使用 ReportDiff 策略时）。

True：完整的差异报告可用
False：差异报告失败或不完整
未知：差异报告挂起

conditions:
- type: ClusterResourcePlacementDiffReported
  status: "True"
  reason: DiffReportComplete
  message: "Configuration differences are reported for all target clusters"
  lastTransitionTime: "2023-11-10T08:16:45Z"
  observedGeneration: 5

了解资源快照

该 observedResourceIndex 字段指示当前正在部署的资源快照：

observedResourceIndex: "1"

当以下情况时，Fleet Manager 会创建资源快照：

资源选择器更改
已修改所选资源

每个快照都有唯一的索引。可以使用以下方法查看快照：

kubectl get clusterresourcesnapshot --selector=kubernetes-fleet.io/resource-index=1

了解每个群集的放置状态

该 placementStatuses 数组包含放置或尝试放置资源的每个群集的详细状态：

placementStatuses:
- clusterName: aks-member-1
  observedResourceIndex: "1"
  conditions:
    - type: ResourceScheduled
      status: "True"
      reason: ScheduleSucceeded
      message: "Successfully scheduled resources for placement in aks-member-1"
      lastTransitionTime: "2023-11-10T08:14:52Z"
      observedGeneration: 5
    - type: RolloutStarted
      status: "True"
      reason: RolloutStarted
      message: "Detected the new changes on the resources and started the rollout process"
      lastTransitionTime: "2023-11-10T08:15:30Z"
      observedGeneration: 5
    - type: Overridden
      status: "True"
      reason: NoOverrideSpecified
      message: "No override rules are configured for the selected resources"
      lastTransitionTime: "2023-11-10T08:15:45Z"
      observedGeneration: 5
    - type: WorkSynchronized
      status: "True"
      reason: AllWorkSynced
      message: "All of the works are synchronized to the latest"
      lastTransitionTime: "2023-11-10T08:16:00Z"
      observedGeneration: 5
    - type: Applied
      status: "True"
      reason: AllWorkHaveBeenApplied
      message: "All corresponding work objects are applied"
      lastTransitionTime: "2023-11-10T08:16:15Z"
      observedGeneration: 5
    - type: Available
      status: "True"
      reason: ResourceAvailable
      message: "All resources are available on the target cluster"
      lastTransitionTime: "2023-11-10T08:16:30Z"
      observedGeneration: 5
  failedPlacements: []
  driftedPlacements: []
  diffedPlacements: []
- clusterName: aks-member-2
  observedResourceIndex: "1"
  conditions:
    - type: ResourceScheduled
      status: "True"
      reason: ScheduleSucceeded
      message: "Successfully scheduled resources for placement in aks-member-2"
      lastTransitionTime: "2023-11-10T08:14:52Z"
      observedGeneration: 5
    - type: Applied
      status: "False"
      reason: AppliedManifestFailedReason
      message: "Failed to apply some manifests"
      lastTransitionTime: "2023-11-10T08:16:15Z"
      observedGeneration: 5
  failedPlacements:
    - kind: Deployment
      name: web-app
      namespace: test
      version: apps/v1
      condition:
        type: Applied
        status: "False"
        reason: AppliedManifestFailedReason
        message: "Failed to apply manifest: insufficient resources"
        lastTransitionTime: "2023-11-10T08:16:15Z"

按群集条件类型

每个群集的状态包括跟踪部署生命周期的条件：

ResourceScheduled

指示是否已成功选择群集进行放置。

- type: ResourceScheduled
  status: "True"
  reason: ScheduleSucceeded
  message: "Successfully scheduled resources for placement in aks-member-1 (affinity score: 0, topology spread score: 0): picked by scheduling policy"
  lastTransitionTime: "2023-11-10T08:14:52Z"
  observedGeneration: 5

RolloutStarted

指示是否已在此特定群集上启动推出。

- type: RolloutStarted
  status: "True"
  reason: RolloutStarted
  message: "Detected new changes on the resources and started the rollout process"
  lastTransitionTime: "2023-11-10T08:15:30Z"
  observedGeneration: 5

重写

指示是否为此群集应用资源替代。

- type: Overridden
  status: "True"
  reason: NoOverrideSpecified
  message: "No override rules are configured for the selected resources"
  lastTransitionTime: "2023-11-10T08:15:45Z"
  observedGeneration: 5

WorkSynchronized

指示是否为此群集创建工作对象。

- type: WorkSynchronized
  status: "True"
  reason: AllWorkSynced
  message: "All of the works are synchronized to the latest"
  lastTransitionTime: "2023-11-10T08:16:00Z"
  observedGeneration: 5

已应用

指示是否成功将所有资源应用于此群集。

- type: Applied
  status: "True"
  reason: AllWorkHaveBeenApplied
  message: "All corresponding work objects are applied"
  lastTransitionTime: "2023-11-10T08:16:15Z"
  observedGeneration: 5

可用

指示所有资源是否都可用，并已在此群集上准备就绪。

- type: Available
  status: "True"
  reason: ResourceAvailable
  message: "All resources are available on the target cluster"
  lastTransitionTime: "2023-11-10T08:16:30Z"
  observedGeneration: 5

DiffReported

指示是否报告此群集的配置差异。

- type: DiffReported
  status: "True"
  reason: DiffReportComplete
  message: "Configuration differences are reported for this cluster"
  lastTransitionTime: "2023-11-10T08:16:45Z"
  observedGeneration: 5

了解失败的位置

如果资源无法应用于群集，则会在 failedPlacements 数组中记录详细信息：

failedPlacements:
- kind: Deployment
  name: my-app
  namespace: default
  version: apps/v1
  condition:
    type: Applied
    status: "False"
    reason: AppliedManifestFailedReason
    message: "Failed to apply manifest: namespaces 'app' not found"
    lastTransitionTime: "2023-12-06T00:09:53Z"
  envelope:
    name: example
    namespace: app
    type: ResourceEnvelope
- kind: Service
  name: my-service
  namespace: default
  version: v1
  condition:
    type: Applied
    status: "False"
    reason: AppliedManifestFailedReason
    message: "Failed to apply manifest: Service 'my-service' is forbidden: User 'system:serviceaccount:fleet-system:fleet-agent' cannot create resource 'services' in API group '' in the namespace 'default'"
    lastTransitionTime: "2023-12-06T00:10:15Z"
- kind: ConfigMap
  name: app-config
  namespace: production
  version: v1
  condition:
    type: Applied
    status: "False"
    reason: AppliedManifestFailedReason
    message: "Failed to apply manifest: configmaps 'app-config' already exists"
    lastTransitionTime: "2023-12-06T00:10:30Z"

每个失败的位置包括：

资源标识：组、版本、类型、名称、命名空间
条件：特定故障条件
信封：信封信息（如果适用）

了解偏移位置

机群管理器始终报告偏离其所需状态的资源：

driftedPlacements:
- kind: Namespace
  name: web
  version: v1
  observationTime: "2025-03-19T06:50:25Z"
  firstDriftedObservedTime: "2025-03-19T06:49:54Z"
  targetClusterObservedGeneration: 12
  observedDrifts:
  - path: "/metadata/labels/owner"
    valueInHub: "simon"
    valueInMember: "chen"
  - path: "/metadata/annotations/purpose"
    valueInHub: "production"
    valueInMember: "testing"
- kind: Deployment
  name: web-app
  namespace: web
  group: apps
  version: v1
  observationTime: "2025-03-19T06:50:25Z"
  firstDriftedObservedTime: "2025-03-19T06:49:54Z"
  targetClusterObservedGeneration: 8
  observedDrifts:
  - path: "/spec/replicas"
    valueInHub: "3"
    valueInMember: "5"
  - path: "/spec/template/spec/containers/0/image"
    valueInHub: "nginx:1.20"
    valueInMember: "nginx:1.21"
- kind: ConfigMap
  name: app-config
  namespace: web
  version: v1
  observationTime: "2025-03-19T06:50:25Z"
  firstDriftedObservedTime: "2025-03-19T06:49:54Z"
  targetClusterObservedGeneration: 5
  observedDrifts:
  - path: "/data/environment"
    valueInHub: "production"
    valueInMember: "staging"

每个偏移位置包括：

资源标识：组、版本、类型、名称、命名空间
观察时间：上次观测偏移时间
firstDriftedObservedTime：首次检测到偏移时
targetClusterObservedGeneration：在成员群集上生成资源
observedDrifts：配置差异的详细列表

了解差异放置

使用 ReportDiff 应用策略时，Fleet Manager 报告配置差异：

diffedPlacements:
- kind: Service
  name: my-service
  namespace: default
  version: v1
  observationTime: "2025-03-19T06:50:25Z"
  firstDiffedObservedTime: "2025-03-19T06:49:54Z"
  targetClusterObservedGeneration: 8
  observedDiffs:
  - path: "/spec/ports/0/nodePort"
    valueInHub: ""
    valueInMember: "30080"
  - path: "/spec/clusterIP"
    valueInHub: ""
    valueInMember: "10.96.100.200"
- kind: Deployment
  name: web-app
  namespace: default
  group: apps
  version: v1
  observationTime: "2025-03-19T06:50:25Z"
  firstDiffedObservedTime: "2025-03-19T06:49:54Z"
  targetClusterObservedGeneration: 12
  observedDiffs:
  - path: "/status/replicas"
    valueInHub: ""
    valueInMember: "3"
  - path: "/status/readyReplicas"
    valueInHub: ""
    valueInMember: "3"
  - path: "/metadata/generation"
    valueInHub: "1"
    valueInMember: "2"

差异放置的结构与偏移放置类似，但用于不同的方案：

偏移位置：在应用资源但随后更改时使用
差异放置：与 ReportDiff 策略一起使用或未满足接管条件时

监视放置进度

若要有效监视放置进度，请检查以下关键指标：

放置的资源：验证 ClusterResourcePlacementWorkSynchronized 是否为 True
总体运行状况：查看 ClusterResourcePlacementApplied 条件
按群集状态：查看每个目标群集的条件
失败的位置：检查数组中的任何 failedPlacements 条目

完成状态示例

下面是显示 ClusterResourcePlacement 的完整状态的综合示例：

apiVersion: placement.kubernetes-fleet.io/v1beta1
kind: ClusterResourcePlacement
metadata:
  name: web-app-placement
  generation: 5
spec:
  resourceSelectors:
  - group: ""
    kind: Namespace
    name: web-app
    version: v1
  - group: apps
    kind: Deployment
    name: web-server
    namespace: web-app
    version: v1
  - group: ""
    kind: Service
    name: web-service
    namespace: web-app
    version: v1
  policy:
    placementType: PickN
    numberOfClusters: 2
    affinity:
      clusterAffinity:
        requiredDuringSchedulingIgnoredDuringExecution:
          clusterSelectorTerms:
          - matchLabels:
              region: chinanorth3
status:
  conditions:
  - type: ClusterResourcePlacementScheduled
    status: "True"
    reason: SchedulingPolicyFulfilled
    message: "found all the clusters needed as specified by the scheduling policy"
    lastTransitionTime: "2023-11-10T08:14:52Z"
    observedGeneration: 5
  - type: ClusterResourcePlacementRolloutStarted
    status: "True"
    reason: RolloutStarted
    message: "All 2 cluster(s) start rolling out the latest resource"
    lastTransitionTime: "2023-11-10T08:15:30Z"
    observedGeneration: 5
  - type: ClusterResourcePlacementOverridden
    status: "True"
    reason: NoOverrideSpecified
    message: "No override rules are configured for the selected resources"
    lastTransitionTime: "2023-11-10T08:15:45Z"
    observedGeneration: 5
  - type: ClusterResourcePlacementWorkSynchronized
    status: "True"
    reason: SynchronizeSucceeded
    message: "All 2 cluster(s) are synchronized to the latest resources on the hub cluster"
    lastTransitionTime: "2023-11-10T08:16:00Z"
    observedGeneration: 5
  - type: ClusterResourcePlacementApplied
    status: "True"
    reason: ApplySucceeded
    message: "The selected resources are successfully applied to 2 clusters"
    lastTransitionTime: "2023-11-10T08:16:15Z"
    observedGeneration: 5
  - type: ClusterResourcePlacementAvailable
    status: "True"
    reason: ResourceAvailable
    message: "The selected resources in 2 cluster are available now"
    lastTransitionTime: "2023-11-10T08:16:30Z"
    observedGeneration: 5
  observedResourceIndex: "1"
  selectedResources:
  - group: ""
    kind: Namespace
    name: web-app
    version: v1
  - group: apps
    kind: Deployment
    name: web-server
    namespace: web-app
    version: v1
  - group: ""
    kind: Service
    name: web-service
    namespace: web-app
    version: v1
  placementStatuses:
  - clusterName: aks-chinanorth3-1
    observedResourceIndex: "1"
    conditions:
    - type: ResourceScheduled
      status: "True"
      reason: ScheduleSucceeded
      message: "Successfully scheduled resources for placement in aks-chinanorth3-1 (affinity score: 0, topology spread score: 0): picked by scheduling policy"
      lastTransitionTime: "2023-11-10T08:14:52Z"
      observedGeneration: 5
    - type: RolloutStarted
      status: "True"
      reason: RolloutStarted
      message: "Detected the new changes on the resources and started the rollout process"
      lastTransitionTime: "2023-11-10T08:15:30Z"
      observedGeneration: 5
    - type: Overridden
      status: "True"
      reason: NoOverrideSpecified
      message: "No override rules are configured for the selected resources"
      lastTransitionTime: "2023-11-10T08:15:45Z"
      observedGeneration: 5
    - type: WorkSynchronized
      status: "True"
      reason: AllWorkSynced
      message: "All of the works are synchronized to the latest"
      lastTransitionTime: "2023-11-10T08:16:00Z"
      observedGeneration: 5
    - type: Applied
      status: "True"
      reason: AllWorkHaveBeenApplied
      message: "All corresponding work objects are applied"
      lastTransitionTime: "2023-11-10T08:16:15Z"
      observedGeneration: 5
    - type: Available
      status: "True"
      reason: ResourceAvailable
      message: "All resources are available on the target cluster"
      lastTransitionTime: "2023-11-10T08:16:30Z"
      observedGeneration: 5
    failedPlacements: []
    driftedPlacements: []
    diffedPlacements: []
  - clusterName: aks-chinanorth3-2
    observedResourceIndex: "1"
    conditions:
    - type: ResourceScheduled
      status: "True"
      reason: ScheduleSucceeded
      message: "Successfully scheduled resources for placement in aks-chinanorth3-2 (affinity score: 0, topology spread score: 0): picked by scheduling policy"
      lastTransitionTime: "2023-11-10T08:14:52Z"
      observedGeneration: 5
    - type: RolloutStarted
      status: "True"
      reason: RolloutStarted
      message: "Detected new changes on the resources and started the rollout process"
      lastTransitionTime: "2023-11-10T08:15:30Z"
      observedGeneration: 5
    - type: Overridden
      status: "True"
      reason: NoOverrideSpecified
      message: "No override rules are configured for the selected resources"
      lastTransitionTime: "2023-11-10T08:15:45Z"
      observedGeneration: 5
    - type: WorkSynchronized
      status: "True"
      reason: AllWorkSynced
      message: "All of the works are synchronized to the latest"
      lastTransitionTime: "2023-11-10T08:16:00Z"
      observedGeneration: 5
    - type: Applied
      status: "True"
      reason: AllWorkHaveBeenApplied
      message: "All corresponding work objects are applied"
      lastTransitionTime: "2023-11-10T08:16:15Z"
      observedGeneration: 5
    - type: Available
      status: "True"
      reason: ResourceAvailable
      message: "All resources are available on the target cluster"
      lastTransitionTime: "2023-11-10T08:16:30Z"
      observedGeneration: 5
    failedPlacements: []
    driftedPlacements: []
    diffedPlacements: []

此示例显示：

成功的计划：放置能够找到两个与放置策略匹配的群集
成功推出：所有资源都部署到两个目标群集
无替代：未配置或需要任何资源替代
同步的工作：创建和同步工作对象
已应用的资源：已成功应用所有资源
可用资源：所有资源都正在运行且可用
清理状态：无失败、偏移或差异放置

若要了解有关资源传播的详细信息，请参阅以下资源：

通过