Alauda DevOps Pipelines Docs
  • 简体中文

    • 配置 Pipeline 组件的资源配额

    目录

    Overview

    配置与 Pipeline 组件相关的资源配额。

    Use Cases

    • 调整 Pipeline 组件的资源配额
    • 设置 TaskRun 创建的 init-containers 和 containers 的默认资源配额

    Prerequisites

    Resource Configuration Guidelines

    在配置资源配额之前：

    • 评估集群的可用资源和容量
    • 考虑工作负载的特性和性能需求
    • 从保守的值开始，根据监控数据进行调整
    • 先在非生产环境中测试配置

    Steps

    Step 1

    编辑 TektonConfig 资源

    $ kubectl edit tektonconfigs.operator.tekton.dev config

    Step 2

    WARNING

    修改配置可能会触发组件 Pods 的滚动更新，可能导致短暂的服务不可用。请在合适的时间执行此操作。

    修改 spec.pipeline.options.deployments 配置示例：

    apiVersion: operator.tekton.dev/v1alpha1
kind: TektonConfig
metadata:
  name: config
spec:
  pipeline:
    options:
      disabled: false

      configMaps:
        config-defaults:
          data:
            # 添加默认容器资源配额
            # 根据集群资源容量和工作负载需求调整以下数值
            default-container-resource-requirements: |
              place-scripts: # 更新 'place-scripts' 容器的资源需求
                requests:
                  memory: "<MEMORY_REQUEST>"  # 例如 "128Mi"
                  cpu: "<CPU_REQUEST>"        # 例如 "250m"
                limits:
                  memory: "<MEMORY_LIMIT>"    # 例如 "512Mi"
                  cpu: "<CPU_LIMIT>"          # 例如 "500m"

              prepare: # 更新 'prepare' 容器的资源需求
                requests:
                  memory: "<MEMORY_REQUEST>"  # 例如 "128Mi"
                  cpu: "<CPU_REQUEST>"        # 例如 "250m"
                limits:
                  memory: "<MEMORY_LIMIT>"    # 例如 "256Mi"
                  cpu: "<CPU_LIMIT>"          # 例如 "500m"

              working-dir-initializer: # 更新 'working-dir-initializer' 容器的资源需求
                requests:
                  memory: "<MEMORY_REQUEST>"  # 例如 "128Mi"
                  cpu: "<CPU_REQUEST>"        # 例如 "250m"
                limits:
                  memory: "<MEMORY_LIMIT>"    # 例如 "512Mi"
                  cpu: "<CPU_LIMIT>"          # 例如 "500m"

              prefix-scripts: # 更新以 'scripts-' 开头的容器的资源需求
                requests:
                  memory: "<MEMORY_REQUEST>"  # 例如 "128Mi"
                  cpu: "<CPU_REQUEST>"        # 例如 "250m"
                limits:
                  memory: "<MEMORY_LIMIT>"    # 例如 "512Mi"
                  cpu: "<CPU_LIMIT>"          # 例如 "500m"

              prefix-sidecar-scripts: # 更新以 'sidecar-scripts-' 开头的容器的资源需求
                requests:
                  memory: "<MEMORY_REQUEST>"  # 例如 "128Mi"
                  cpu: "<CPU_REQUEST>"        # 例如 "250m"
                limits:
                  memory: "<MEMORY_LIMIT>"    # 例如 "512Mi"
                  cpu: "<CPU_LIMIT>"          # 例如 "500m"

              sidecar-tekton-log-results: # 更新 'sidecar-tekton-log-results' 容器的资源需求
                requests:
                  memory: "<MEMORY_REQUEST>"  # 例如 "128Mi"
                  cpu: "<CPU_REQUEST>"        # 例如 "100m"
                limits:
                  memory: "<MEMORY_LIMIT>"    # 例如 "256Mi"
                  cpu: "<CPU_LIMIT>"          # 例如 "250m"

      deployments:
        # 根据集群容量和性能需求调整以下资源数值
        tekton-pipelines-controller:
          spec:
            replicas: <REPLICA_COUNT>  # 例如 1
            template:
              spec:
                containers:
                  - name: tekton-pipelines-controller
                    resources:
                      requests:
                        cpu: <CPU_REQUEST>        # 例如 "500m"
                        memory: <MEMORY_REQUEST>  # 例如 "512Mi"
                      limits:
                        cpu: <CPU_LIMIT>          # 例如 "1"
                        memory: <MEMORY_LIMIT>    # 例如 "1Gi"

        tekton-pipelines-remote-resolvers:
          spec:
            replicas: <REPLICA_COUNT>  # 例如 1
            template:
              spec:
                containers:
                  - name: controller
                    resources:
                      requests:
                        cpu: <CPU_REQUEST>        # 例如 "200m"
                        memory: <MEMORY_REQUEST>  # 例如 "256Mi"
                      limits:
                        cpu: <CPU_LIMIT>          # 例如 "500m"
                        memory: <MEMORY_LIMIT>    # 例如 "512Mi"

        tekton-pipelines-webhook:
          spec:
            replicas: <REPLICA_COUNT>  # 例如 1
            template:
              spec:
                containers:
                  - name: webhook
                    resources:
                      requests:
                        cpu: <CPU_REQUEST>        # 例如 "500m"
                        memory: <MEMORY_REQUEST>  # 例如 "256Mi"
                      limits:
                        cpu: <CPU_LIMIT>          # 例如 "1"
                        memory: <MEMORY_LIMIT>    # 例如 "500Mi"

        tekton-events-controller:
          spec:
            replicas: <REPLICA_COUNT>  # 例如 1
            template:
              spec:
                containers:
                  - name: tekton-events-controller
                    resources:
                      requests:
                        cpu: <CPU_REQUEST>        # 例如 "100m"
                        memory: <MEMORY_REQUEST>  # 例如 "100Mi"
                      limits:
                        cpu: <CPU_LIMIT>          # 例如 "200m"
                        memory: <MEMORY_LIMIT>    # 例如 "256Mi"

    Step 3

    提交配置并等待 Pods 更新。

    $ kubectl get pods -n tekton-pipelines -w

NAME                                                    READY   STATUS    RESTARTS   AGE
tekton-pipelines-controller-648d87488b-fq9bc            1/1     Running   0          2m21s
tekton-pipelines-remote-resolvers-79554f5959-cbm6x      1/1     Running   0          2m21s
tekton-pipelines-webhook-5cd9847998-864zf               1/1     Running   0          2m20s
tekton-events-controller-5c97b7554c-m59m6               1/1     Running   0          2m21s

    Operation Result

    您可以看到与 Pipeline 相关组件的资源配额配置已生效。

    $ kubectl get deployments.apps -n tekton-pipelines tekton-pipelines-controller tekton-pipelines-remote-resolvers tekton-pipelines-webhook tekton-events-controller -o yaml | grep 'resources:' -A 6

          resources:
            limits:
              cpu: "1"
              memory: 1Gi
            requests:
              cpu: 500m
              memory: 512Mi
--
          resources:
            limits:
              cpu: 500m
              memory: 512Mi
            requests:
              cpu: 200m
              memory: 256Mi
--
          resources:
            limits:
              cpu: "2"
              memory: 500Mi
            requests:
              cpu: "1"
              memory: 256Mi
--
          resources:
            limits:
              cpu: 200m
              memory: 256Mi
            requests:
              cpu: 100m
              memory: 100Mi

    验证 Pod 资源配额配置

    创建一个 TaskRun

    $ cat <<'EOF' | kubectl create -f -
apiVersion: tekton.dev/v1
kind: TaskRun
metadata:
  name: hello
  namespace: default
spec:
  taskSpec:
    steps:
      - name: hello
        image: alpine
        command: ["echo", "hello"]
EOF

    等待 TaskRun 完成

    $ kubectl get taskruns.tekton.dev -n default hello

NAME    SUCCEEDED   REASON      STARTTIME   COMPLETIONTIME
hello   True        Succeeded   2m41s       2m28s

    查看 Pod 资源配额配置

    $ kubectl get pods -n default hello-hello-pod -o yaml

apiVersion: v1
kind: Pod
metadata:
  name: hello-pod
  namespace: default
spec:
  containers:
    - image: alpine
      name: step-hello
      resources: {}
  initContainers:
    - name: prepare
      resources:
        limits:
          cpu: 100m
          memory: 256Mi
        requests:
          cpu: 50m
          memory: 64Mi

    您可以看到 Pod 中 initContainers 容器 prepare 的资源配额与 config-defaults ConfigMap 中配置的资源配额一致。

    References