Configure Kubernetes Connector resource limits

The Kubernetes Connector consists of multiple components that require CPU and memory allocation. Follow these steps to set limits that align with your resource needs.

1. Download the values file

   From the Kubernetes Connect wizard, click Generate to download the K8s-security-profile-panw-DATE.values.yaml file from the first step.
2. Adjust workload requests and limits Edit the file to adjust the requests and limits of the workloads. Use this example as a reference when configuring your file:

konnector-template:
  enabled: true

# ============================================================================
# Core — Cortex Manager (k8sConnector) — full override
# ============================================================================
coreValues:
  resources:
    k8sConnector:
      overrides:
        requests:
          cpu: "300m"
          memory: "256Mi"
        limits:
          cpu: "500m"
          memory: "512Mi"

# ============================================================================
# Monitoring — OpenTelemetry Collector (otel) — full override
# ============================================================================
monitoringValues:
  enabled: true
  resources:
    otel:
      overrides:
        requests:
          cpu: "100m"
          memory: "200Mi"
        limits:
          cpu: "300m"
          memory: "400Mi"

# ============================================================================
# Posture — admission controller, orchestrator, and spawned jobs
# ============================================================================
postureValues:
  enabled: true

  admissionController:
    enabled: true
    failurePolicy: Ignore

  jobOrchestrator:
    scanIntervalHours: 12

  resources:
    # Admission Controller — full override
    admissionController:
      overrides:
        requests:
          cpu: "100m"
          memory: "300Mi"
        limits:
          cpu: "300m"
          memory: "600Mi"

    # Job Orchestrator — CPU request + CPU limit only
    # (memory keeps defaults: requests.memory=75Mi, limits.memory=100Mi)
    jobOrchestrator:
      overrides:
        requests:
          cpu: "50m" # was 25m
        limits:
          cpu: "100m" # was 50m

    # Resources for jobs spawned by the orchestrator
    jobResources:
      # Host Scanner — memory limit only
      # (CPU and memory request keep defaults)
      host-scanner:
        overrides:
          limits:
            memory: "8Gi" # was 4.5Gi

      # Inventory Collector — CPU request + CPU limit only
      # (memory keeps defaults: requests.memory=50Mi, limits.memory=100Mi)
      inventory-collector:
        overrides:
          requests:
            cpu: "50m" # was 25m
          limits:
            cpu: "100m" # was 50m

      # Compliance K8s Scanner — full override
      compliance-k8s-scanner:
        overrides:
          requests:
            cpu: "300m"
            memory: "100Mi"
          limits:
            cpu: "750m"
            memory: "200Mi"

      # Scan Result Publisher — full override
      scan-result-publisher:
        overrides:
          requests:
            cpu: "50m"
            memory: "100Mi"
          limits:
            cpu: "100m"
            memory: "750Mi"

      # Node VM Discovery — full override
      node-vm-discovery:
        overrides:
          requests:
            cpu: "30m"
            memory: "100Mi"
          limits:
            cpu: "100m"
            memory: "200Mi"

1. Deploy the Kubernetes Connector Deploy the Kubernetes Connector using Helm:

helm upgrade --install konnector oci://us-central1-docker.pkg.dev/qa2-test-YYY/agent docker/helm/konnector-launcher \
  --wait-for-jobs \
  --create-namespace \
  --namespace panw \
  --values K8s-security-profile-panw-DATE.values.yaml \
  --set-file konnector-upgrader.rawValuesContent=K8s-security-profile-panw-DATE.values.yaml