> ## Documentation Index
> Fetch the complete documentation index at: https://docs.costgraph.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Configuration

> Helm values reference for the CostGraph Operator

The CostGraph Operator is configured through Helm `values.yaml`. The defaults work for most clusters; this page covers the options you're most likely to change.

## Global settings

<ParamField body="global.clusterName" type="string" required>
  Name used to identify this cluster in the CostGraph platform.
</ParamField>

<ParamField body="global.apiKey" type="string">
  API key for authenticating with CostGraph. Don't commit a real key — prefer `existingSecret` in production.
</ParamField>

<ParamField body="global.existingSecret" type="string">
  Name of a pre-existing Secret holding the API key. When set, `apiKey` is ignored.
</ParamField>

<ParamField body="global.existingSecretKey" type="string" default="apiKey">
  Key within `existingSecret` that holds the API key value.
</ParamField>

<ParamField body="global.backendURL" type="string" default="https://api.costgraph.ai">
  Base URL of the CostGraph API.
</ParamField>

<ParamField body="global.namespace" type="string" default="costgraph">
  Namespace the operator runs in.
</ParamField>

## Resource exporter

`operatorKubernetes` watches cluster resources and exports them to CostGraph.

<ParamField body="operatorKubernetes.enabled" type="boolean" default="true">
  Deploy the resource exporter.
</ParamField>

<ParamField body="operatorKubernetes.replicas" type="number" default="1">
  Keep at `1` — the exporter is not designed to run multiple replicas.
</ParamField>

<ParamField body="operatorKubernetes.resources" type="object">
  Resource requests and limits.
</ParamField>

<ParamField body="operatorKubernetes.nodeSelector / tolerations / affinity" type="object">
  Standard pod scheduling controls.
</ParamField>

<ParamField body="operatorKubernetes.env" type="array">
  Additional environment variables.
</ParamField>

## Operational/Compute Metrics scraper

`operatorPrometheus` scrapes in-cluster metrics endpoints and remote-writes them to CostGraph.

<ParamField body="operatorPrometheus.enabled" type="boolean" default="true">
  Deploy the operational/compute metrics scraper.
</ParamField>

<ParamField body="operatorPrometheus.config.remoteWriteURL" type="string" default="https://tsdb.costgraph.ai">
  Remote-write endpoint for collected metrics.
</ParamField>

<ParamField body="operatorPrometheus.config.remoteWritePath" type="string" default="/api/v1/write">
  Remote-write path appended to the URL.
</ParamField>

<ParamField body="operatorPrometheus.config.prometheusAPIProvider" type="string" default="prometheus">
  Metrics API flavor. `prometheus` or `victoriametrics`.
</ParamField>

<ParamField body="operatorPrometheus.config.httpTimeout" type="string" default="30s">
  Timeout for scrape and backend requests.
</ParamField>

## AI gateway scraper

`aiGatewayScraper` reads token usage from AI gateways running in the cluster and
remote-writes it to the short-retention tier. It is a second instance of the
same scraper image; the difference is `remoteWritePath`, because raw gateway
series are high cardinality and are not intended to live as long as the
operator's scraped metrics.

<ParamField body="aiGatewayScraper.enabled" type="boolean" default="false">
  Deploy the AI gateway scraper.
</ParamField>

<ParamField body="aiGatewayScraper.config.remoteWritePath" type="string" default="/api/v1/write/short">
  Short-retention tier. Recording rules distil these series before anything
  reaches 90-day retention.
</ParamField>

It takes **no** `scrapeTargets`. Targets are served by CostGraph from the
connections you create in **Settings -> Integrations**, fetched at start and
refreshed while it runs, so connecting a gateway never edits these values. See
[AI gateways](/costgraph/integrations/ai-gateways).

A target you define yourself in `operatorPrometheus.scrapeTargets` always wins
over a served one, and a served target is only applied when it carries a
metric-name filter.

## Network Metrics scraper

`flowtrace` scrapes per-container network metrics and, optionally, host-process network metrics, then remote-writes them to CostGraph.

<Note>
  The DaemonSet mounts the host filesystems at:

  * **cgroupfs** — `/host/sys/fs/cgroup`.
  * **procfs** — `/host/proc`.
</Note>

<ParamField body="flowtrace.enabled" type="boolean" default="true">
  Deploy the network metrics scraper.
</ParamField>

<ParamField body="flowtrace.config.remoteWriteURL" type="string" default="https://tsdb.costgraph.ai">
  Remote-write endpoint for collected metrics.
</ParamField>

<ParamField body="flowtrace.config.remoteWritePath" type="string" default="/api/v1/write/short">
  Remote-write path appended to the URL.
</ParamField>

<ParamField body="flowtrace.config.httpTimeout" type="string" default="30s">
  Timeout for write and backend requests.
</ParamField>

<ParamField body="flowtrace.config.includeHostFlows" type="boolean" default="false">
  Container network flows are collected by default. Host-process network flows are collected only when this is enabled.
</ParamField>

### Scrape targets

`operatorPrometheus.scrapeTargets` is a map of the in-cluster `/metrics` endpoints to scrape. The chart ships defaults for cAdvisor, kube-state-metrics, node-exporter, and the kubelet (DCGM is off). Add a source by appending an entry — no new operator image is required.

Each entry accepts:

<ParamField body="enabled" type="boolean" default="true">
  Include this target.
</ParamField>

<ParamField body="kind" type="string" default="pod">
  `pod` discovers and scrapes matching pods; `node` scrapes each node's kubelet through the API-server proxy.
</ParamField>

<ParamField body="namespace" type="string">
  Pod targets only — namespace to discover pods in.
</ParamField>

<ParamField body="labelSelector" type="string">
  Label selector for pod (or node) discovery. Required for pod targets.
</ParamField>

<ParamField body="port" type="number">
  Pod targets only — port the exporter listens on.
</ParamField>

<ParamField body="path" type="string" default="/metrics">
  Metrics path to scrape.
</ParamField>

<ParamField body="nodeMetrics" type="boolean" default="false">
  Stamp `costgraph_instance_hostname` and `costgraph_instance_id` on node-scoped sources like cAdvisor and node-exporter.
</ParamField>

<ParamField body="metricNamePrefixes" type="array">
  Keep only metric families whose names match one of these prefixes. Empty keeps everything.
</ParamField>

## Bundled exporters

The chart pulls in the metrics sources as subcharts. Disable any you already run in the cluster. Disable `\flowtrace` if you do not need network tracing.

<ParamField body="kube-state-metrics.enabled" type="boolean" default="true">
  Kubernetes object-level metrics (pods, deployments, nodes).
</ParamField>

<ParamField body="cadvisor.enabled" type="boolean" default="true">
  Per-container resource usage. Uses the `baselinehq/cadvisor` image.
</ParamField>

<ParamField body="prometheus-node-exporter.enabled" type="boolean" default="true">
  Host-level metrics (CPU, memory, disk, network) on port `9100`.
</ParamField>

<ParamField body="dcgm-exporter.enabled" type="boolean" default="false">
  NVIDIA GPU metrics. Enable only on clusters with GPU nodes; restrict it to GPU nodes with `dcgm-exporter.nodeSelector` and `dcgm-exporter.tolerations`.
</ParamField>

<ParamField body="flowtrace.enabled" type="boolean" default="true">
  Per-container and, optionally, host per-process network metrics (bytes sent/received, packets sent/received, connection count).
</ParamField>

## Full values reference

<Accordion title="Complete values.yaml">
  ```yaml theme={null}
  global:
    clusterName: ""
    apiKey: ""
    existingSecret: ""
    existingSecretKey: ""
    backendURL: "https://api.costgraph.ai"
    namespace: "costgraph"

  serviceAccount:
    create: true
    automount: true

  kube-state-metrics:
    enabled: true

  cadvisor:
    enabled: true
    image:
      registry: ghcr.io
      repository: baselinehq/cadvisor
      tag: "0.56.3-baseline"

  prometheus-node-exporter:
    enabled: true
    service:
      port: 9100
      targetPort: 9100

  dcgm-exporter:
    enabled: false
    nodeSelector:
      accelerator: nvidia
    tolerations: []

  operatorKubernetes:
    enabled: true
    replicas: 1
    resources:
      requests:
        cpu: 10m
        memory: 64Mi
    nodeSelector: {}
    tolerations: []
    affinity: {}
    env: []

  operatorPrometheus:
    enabled: true
    replicas: 1
    config:
      remoteWriteURL: "https://tsdb.costgraph.ai"
      remoteWritePath: "/api/v1/write"
      prometheusAPIProvider: "prometheus"
      httpTimeout: "30s"
    scrapeTargets:
      cadvisor:
        enabled: true
        labelSelector: "app.kubernetes.io/name=cadvisor"
        port: 8080
        nodeMetrics: true
      kube-state-metrics:
        enabled: true
        labelSelector: "app.kubernetes.io/name=kube-state-metrics"
        port: 8080
      node-exporter:
        enabled: true
        labelSelector: "app.kubernetes.io/name=prometheus-node-exporter"
        port: 9100
        nodeMetrics: true
      kubelet:
        enabled: true
        kind: node
        nodeMetrics: true
        metricNamePrefixes:
          - kubelet_
      dcgm-exporter:
        enabled: false
        labelSelector: "app.kubernetes.io/name=dcgm-exporter"
        port: 9400
        nodeMetrics: true
    resources:
      requests:
        cpu: 100m
        memory: 64Mi
    env: []

  # flowtrace runs on every node and tolerates all taints by default, 
  # requires BTF and cgroup v2, and enters CrashLoopBackOff on incompatible nodes. 
  # Use nodeSelector to target compatible nodes in mixed clusters and
  # set flowtrace.enabled to false when no compatible nodes exist.

  flowtrace:
    enabled: true
    config:
      remoteWriteURL: "https://tsdb.costgraph.ai"
      remoteWritePath: "/api/v1/write/short"
      pushInterval: "4m"
      includeHostFlows: false
      httpTimeout: "30s"
    maxUnavailable: 1
    priorityClassName: ""
    resources:
      requests:
        cpu: 50m
        memory: 64Mi
      limits:
        memory: 256Mi
    tolerations:
      - operator: Exists
    nodeSelector: {}
    affinity: {}
    podAnnotations: {}
    podLabels: {}
    env: []
  ```
</Accordion>

## Node and volume pricing

Nodes and volumes outside a supported cloud are priced with a
`costgraph.ai/pricing-id` label. See [On-prem pricing](/costgraph/operator/on-prem-pricing).