0

有谁知道为什么我从 Openshift Operator 安装 3Scale APIManager 需要这么多资源？

默认安装要求 4 个 CPU 和大约 9Gi 的内存，但有 90Gi 的限制。所以它不会安装，除非我的项目分配了超过 90Gi 的内存。

如果我使用 100 个线程进行负载测试，我最多可以在所有 pod 上加载 2 个 CPU 和 4Gi 内存。

这是我的 API 管理器 yaml

apiVersion: apps.3scale.net/v1alpha1
kind: APIManager
metadata:
  annotations:
    apps.3scale.net/apimanager-threescale-version: '2.9'
    apps.3scale.net/threescale-operator-version: 0.6.0
  name: apimanager
  generation: 2
  namespace: user-greg-clinker-sandbox <--TODO
spec:
  imageStreamTagImportInsecure: false
  resourceRequirementsEnabled: true
  system:
    appSpec:
      replicas: 1
    database:
      postgresql:
        persistentVolumeClaim:
          storageClassName: nfs-non-vdi-retain-backup-enc
    fileStorage:
      persistentVolumeClaim:
        storageClassName: nfs-non-vdi-retain-backup-enc
    redisPersistentVolumeClaim:
      storageClassName: nfs-non-vdi-retain-backup-enc
    sidekiqSpec:
      replicas: 1
    sphinxSpec: {}
  appLabel: 3scale-api-management
  zync:
    appSpec:
      replicas: 1
    queSpec:
      replicas: 1
  backend:
    cronSpec:
      replicas: 1
    listenerSpec:
      replicas: 1
    redisPersistentVolumeClaim:
      storageClassName: nfs-non-vdi-retain-backup-enc
    workerSpec:
      replicas: 1
  tenantName: 3scale
  apicast:
    managementAPI: status
    openSSLVerify: false
    productionSpec:
      replicas: 3
    registryURL: 'http://apicast-staging:8090/policies
    responseCodes: true
    stagingSpec:
      replicas: 3
  wildcardDomain: 3scale2.apps.ocp.net

这是我之后的资源使用情况

spec:
  hard:
    limits.cpu: '24'
    limits.memory: 128Gi
    requests.cpu: '6'
    requests.memory: 13743895347200m
  scopes:
    - NotTerminating
status:
  hard:
    limits.cpu: '24'
    limits.memory: 128Gi
    requests.cpu: '6'
    requests.memory: 13743895347200m
  used:
    limits.cpu: 16100m
    limits.memory: 85048677Ki
    requests.cpu: 3950m
    requests.memory: '8832423808'
4

0 回答 0