values.yaml

# Copyright SAP SE
# SPDX-License-Identifier: Apache-2.0

owner-info:
  enabled: true
  helm-chart-url: "https://github.com/cobaltcore-dev/cortex/helm/bundles/cortex-nova"
  maintainers:
    - "arno.uhlig@sap.com"
    - "julius.clausnitzer@sap.com"
    - "malte.viering@sap.com"
    - "marcel.gute@sap.com"
    - "markus.wieland@sap.com"
    - "p.matthes@sap.com"
  support-group: "workload-management"
  service: "cortex-nova"

alerts:
  enabled: true
  prometheus: openstack

serviceMonitor:
  extraLabels: {}

# SSO certificate to use.
sharedSSOCert: &sharedSSOCert
  # Certificate "public key". (Optional, remove this key if not needed)
  cert: |
    -----BEGIN CERTIFICATE-----
    Your certificate here
    -----END CERTIFICATE-----
  # Certificate private key. (Optional, remove this key if not needed)
  key: |
    -----BEGIN PRIVATE KEY-----
    Your private key here
    -----END PRIVATE KEY-----
  # Whether the certificate is self-signed.
  # If true, the certificate is not verified.
  selfSigned: "false"

postgres:
  host: cortex-nova-postgresql
  user: postgres
  password: secret
  database: postgres
  port: "5432"

prometheus:
  url: "https://path-to-your-prometheus"
  sso:
    enabled: false
    <<: *sharedSSOCert

openstack:
  url: "https://path-to-keystone/v3"
  availability: public
  username: openstack-user-with-all-project-read-access
  password: openstack-user-password
  projectName: openstack-project-of-user
  userDomainName: openstack-domain-of-user
  projectDomainName: openstack-domain-of-project-scoped-to
  sso:
    enabled: false
    <<: *sharedSSOCert

kvm:
  # Use this flag to enable/disable KVM host related features.
  enabled: false
  # Params for the filter_requested_destination filter step.
  filterRequestedDestinationParams: []
  # Example:
  # filterRequestedDestinationParams:
  #   - {key: ignoredHostnames, stringListValue: []}
  #   - {key: ignoredAggregates, stringListValue: []}

cortex: &cortex
  crd: {enable: false}
  # Disable the default ServiceMonitor and metrics service from the kubebuilder stack.
  # A centralized ServiceMonitor and metrics service are defined in the cortex-nova chart.
  metrics: {enable: false}
  prometheus: {enable: false}
  conf: &cortexConf
    schedulingDomain: nova
    apiservers:
      home:
        gvks:
          - cortex.cloud/v1alpha1/Decision
          - cortex.cloud/v1alpha1/DecisionList
          - cortex.cloud/v1alpha1/Descheduling
          - cortex.cloud/v1alpha1/DeschedulingList
          - cortex.cloud/v1alpha1/Pipeline
          - cortex.cloud/v1alpha1/PipelineList
          - cortex.cloud/v1alpha1/Knowledge
          - cortex.cloud/v1alpha1/KnowledgeList
          - cortex.cloud/v1alpha1/Datasource
          - cortex.cloud/v1alpha1/DatasourceList
          - cortex.cloud/v1alpha1/KPI
          - cortex.cloud/v1alpha1/KPIList
          - cortex.cloud/v1alpha1/Reservation
          - cortex.cloud/v1alpha1/ReservationList
          - kvm.cloud.sap/v1/Hypervisor
          - kvm.cloud.sap/v1/HypervisorList
          - v1/Secret
    keystoneSecretRef:
      name: cortex-nova-openstack-keystone
      namespace: default
    # ssoSecretRef:
    #   name: cortex-nova-openstack-sso
    #   namespace: default
    monitoring:
      labels: &cortexMonitoringLabels
        github_org: cobaltcore-dev
        github_repo: cortex

# Split deployments
cortex-scheduling-controllers:
  <<: *cortex
  namePrefix: cortex-nova-scheduling
  rbac:
    # The cortex nova scheduling controllers need hypervisor crd access.
    hypervisor: {enable: true}
  # Enable webhook that will validate CRDs for the scheduling controllers.
  webhook: {enable: true}
  certmanager: {enable: true} # Needed for the webhook TLS certificates.
  conf:
    <<: *cortexConf
    leaderElectionID: cortex-nova-scheduling
    monitoring:
      labels:
        <<: *cortexMonitoringLabels
        component: nova-scheduling
    enabledControllers:
      - nova-pipeline-controllers
      - nova-deschedulings-executor
      - hypervisor-overcommit-controller
      - explanation-controller
      - committed-resource-reservations-controller
      - failover-reservations-controller
    enabledTasks:
      - nova-decisions-cleanup-task
    # CommittedResourceFlavorGroupPipelines maps flavor group IDs to pipeline names for CR reservations
    # This allows different scheduling strategies per flavor group (e.g., HANA vs GP)
    committedResourceFlavorGroupPipelines:
      "2152": "kvm-hana-bin-packing"  # HANA flavor group
      "2101": "kvm-general-purpose-load-balancing"  # General Purpose flavor group
      "*": "kvm-general-purpose-load-balancing"  # Catch-all fallback
    # Default pipeline for CR reservations when no CommittedResourceFlavorGroupPipelines entry matches
    committedResourcePipelineDefault: "kvm-general-purpose-load-balancing"
    # How often to re-verify active reservations
    # 5m = 300000000000 nanoseconds
    committedResourceRequeueIntervalActive: 300000000000
    # How often to retry when knowledge is not ready
    # 1m = 60000000000 nanoseconds
    committedResourceRequeueIntervalRetry: 60000000000
    # Timeout for watching reservations to become ready before rolling back
    # 10s = 10000000000 nanoseconds
    committedResourceChangeAPIWatchReservationsTimeout: 10000000000
    # How often to poll reservation status during watch
    # 500ms = 500000000 nanoseconds
    committedResourceChangeAPIWatchReservationsPollInterval: 500000000
    # Whether the change-commitments API endpoint is active
    # When false, the endpoint returns HTTP 503. The info endpoint remains available.
    committedResourceEnableChangeCommitmentsAPI: true
    # OvercommitMappings is a list of mappings that map hypervisor traits to
    # overcommit ratios. Note that this list is applied in order, so if there
    # are multiple mappings applying to the same hypervisors, the last mapping
    # in this list will override the previous ones.
    overcommitMappings: []
    # Failover reservations controller configuration
    # Name of the Datasource CRD that provides database connection info for Nova VM data
    datasourceName: nova-servers
    # URL of the nova external scheduler API for placement decisions
    schedulerURL: "http://localhost:8080/scheduler/nova/external"
    # Maps flavor name patterns (glob) to required failover count
    # Example: {"hana_*": 2, "m1.xlarge": 1}
    flavorFailoverRequirements:
      "*": 1
    # How often to check for missing failover reservations (periodic bulk reconciliation)
    reconcileInterval: 15m
    # Used when maxVMsToProcess limits processing, allows faster catch-up and for the first reconcile
    shortReconcileInterval: 1m
    # Number of max VMs to process in one periodic reconciliation loop
    maxVMsToProcess: 5
    # Minimum successful reservations to use short interval
    minSuccessForShortInterval: 1
    # Maximum failures allowed to still use short interval
    maxFailuresForShortInterval: 99
    # If true, uses hypervisor CRD as source of truth for VM location instead of postgres
    trustHypervisorLocation: true
    # How often to re-validate acknowledged failover reservations
    revalidationInterval: 30m
    # Prevents creating multiple new reservations on the same hypervisor per cycle
    limitOneNewReservationPerHypervisor: false

cortex-knowledge-controllers:
  <<: *cortex
  namePrefix: cortex-nova-knowledge
  rbac:
    # The cortex nova scheduling controllers need hypervisor crd access.
    hypervisor: {enable: true}
  conf:
    <<: *cortexConf
    leaderElectionID: cortex-nova-knowledge
    monitoring:
      labels:
        <<: *cortexMonitoringLabels
        component: nova-knowledge
    enabledControllers:
      - datasource-controllers
      - knowledge-controllers
      - kpis-controller
    # How often the commitments syncer reconciles Limes commitments to Reservation CRDs
    # 1h = 3600000000000 nanoseconds
    committedResourceSyncInterval: 3600000000000
    enabledTasks:
      - commitments-sync-task

# Custom configuration for the cortex postgres chart.
cortex-postgres:
  fullnameOverride: cortex-nova-postgresql