kube-prometheus-example.yaml

# 20220531 Example values file for Bitnami kube-prometheus Helm chart installation (v8.0.0)
# Refer to https://github.com/bitnami/charts/tree/master/bitnami/kube-prometheus

prometheus:
  # Queries can take ~> 10 minutes on a large busy cluster. Allow very long queries to avoid timeouts.
  # See https://github.com/prometheus-operator/prometheus-operator/blob/master/Documentation/api.md#queryspec
  querySpec:
    timeout: 1800s

  ## How long to retain metrics
  retention: 120d

  ## Maximum size of metrics
  retentionSize: ""

  # For a very large cluster (~7000 cores), to process large volumes of job metrics and accounting records.
  resources:
    requests:
      cpu: "2000m"
      memory: "16Gi"
    limits:
      cpu: "8000m"
      memory: "64Gi"

  ## for persistent Prometheus data. Make sure to review https://github.com/bitnami/charts/tree/master/bitnami/kube-prometheus#prometheus-parameters
  ## to ensure the storage configuration is appropriate for your cluster.
  persistence:
    enabled: true

kube-state-metrics:
  # Limit metrics collection to only the namespace where compute jobs are executed.
  namespaces: "harvester"

  ## Collectors to be enabled. Only pods are required for accounting.
  kubeResources:
    certificatesigningrequests: false
    configmaps: false
    cronjobs: false
    daemonsets: false
    deployments: false
    endpoints: false
    horizontalpodautoscalers: false
    ingresses: false
    jobs: false
    limitranges: false
    mutatingwebhookconfigurations: false
    namespaces: false
    networkpolicies: false
    nodes: false
    persistentvolumeclaims: false
    persistentvolumes: false
    poddisruptionbudgets: false
    pods: true
    replicasets: false
    replicationcontrollers: false
    resourcequotas: false
    secrets: false
    services: false
    statefulsets: false
    storageclasses: false
    verticalpodautoscalers: false
    validatingwebhookconfigurations: false
    volumeattachments: false

  # According to https://github.com/kubernetes/kube-state-metrics/blob/master/README.md 
  # this should be more than enough for a cluster with 500 nodes.
  resources:
    limits: 
      cpu: 2000m
      memory: 2048Mi
    requests: 
      cpu: 200m
      memory: 256Mi

alertmanager:
  enabled: false