1. Documentation
    2. history bug_report picture_as_pdf
×

Configuring metrics for the monitoring stack

As a cluster administrator, you can configure the OpenTelemetry Collector custom resource (CR) to perform the following tasks:

  • Create a Prometheus ServiceMonitor CR for scraping the Collector’s pipeline metrics and the enabled Prometheus exporters.

  • Configure the Prometheus receiver to scrape metrics from the in-cluster monitoring stack.

Configuration for sending metrics to the monitoring stack

You can configure the OpenTelemetryCollector custom resource (CR) to create a Prometheus ServiceMonitor CR or a PodMonitor CR for a sidecar deployment. A ServiceMonitor can scrape Collector’s internal metrics endpoint and Prometheus exporter metrics endpoints.

Example of the OpenTelemetry Collector CR with the Prometheus exporter
apiVersion: opentelemetry.io/v1alpha1
kind: OpenTelemetryCollector
spec:
  mode: deployment
  observability:
    metrics:
      enableMetrics: true (1)
  config: |
    exporters:
      prometheus:
        endpoint: 0.0.0.0:8889
        resource_to_telemetry_conversion:
          enabled: true # by default resource attributes are dropped
    service:
      telemetry:
        metrics:
          address: ":8888"
      pipelines:
        metrics:
          receivers: [otlp]
          exporters: [prometheus]
1 Configures the Red Hat build of OpenTelemetry Operator to create the Prometheus ServiceMonitor CR or PodMonitor CR to scrape the Collector’s internal metrics endpoint and the Prometheus exporter metrics endpoints.

Setting enableMetrics to true creates the following two ServiceMonitor instances:

  • One ServiceMonitor instance for the <instance_name>-collector-monitoring service. This ServiceMonitor instance scrapes the Collector’s internal metrics.

  • One ServiceMonitor instance for the <instance_name>-collector service. This ServiceMonitor instance scrapes the metrics exposed by the Prometheus exporter instances.

Alternatively, a manually created Prometheus PodMonitor CR can provide fine control, for example removing duplicated labels added during Prometheus scraping.

Example of the PodMonitor CR that configures the monitoring stack to scrape the Collector metrics
apiVersion: monitoring.coreos.com/v1
kind: PodMonitor
metadata:
  name: otel-collector
spec:
  selector:
    matchLabels:
      app.kubernetes.io/name: <cr_name>-collector (1)
  podMetricsEndpoints:
  - port: metrics (2)
  - port: promexporter (3)
    relabelings:
    - action: labeldrop
      regex: pod
    - action: labeldrop
      regex: container
    - action: labeldrop
      regex: endpoint
    metricRelabelings:
    - action: labeldrop
      regex: instance
    - action: labeldrop
      regex: job
1 The name of the OpenTelemetry Collector CR.
2 The name of the internal metrics port for the OpenTelemetry Collector. This port name is always metrics.
3 The name of the Prometheus exporter port for the OpenTelemetry Collector.

Configuration for receiving metrics from the monitoring stack

A configured OpenTelemetry Collector custom resource (CR) can set up the Prometheus receiver to scrape metrics from the in-cluster monitoring stack.

Example of the OpenTelemetry Collector CR for scraping metrics from the in-cluster monitoring stack
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: otel-collector
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: cluster-monitoring-view (1)
subjects:
  - kind: ServiceAccount
    name: otel-collector
    namespace: observability
---
kind: ConfigMap
apiVersion: v1
metadata:
  name: cabundle
  namespce: observability
  annotations:
    service.beta.openshift.io/inject-cabundle: "true" (2)
---
apiVersion: opentelemetry.io/v1alpha1
kind: OpenTelemetryCollector
metadata:
  name: otel
  namespace: observability
spec:
  volumeMounts:
    - name: cabundle-volume
      mountPath: /etc/pki/ca-trust/source/service-ca
      readOnly: true
  volumes:
    - name: cabundle-volume
      configMap:
        name: cabundle
  mode: deployment
  config: |
    receivers:
      prometheus: (3)
        config:
          scrape_configs:
            - job_name: 'federate'
              scrape_interval: 15s
              scheme: https
              tls_config:
                ca_file: /etc/pki/ca-trust/source/service-ca/service-ca.crt
              bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token
              honor_labels: false
              params:
                'match[]':
                  - '{__name__="<metric_name>"}' (4)
              metrics_path: '/federate'
              static_configs:
                - targets:
                  - "prometheus-k8s.openshift-monitoring.svc.cluster.local:9091"
    exporters:
      debug: (5)
        verbosity: detailed
    service:
      pipelines:
        metrics:
          receivers: [prometheus]
          processors: []
          exporters: [debug]
1 Assigns the cluster-monitoring-view cluster role to the service account of the OpenTelemetry Collector so that it can access the metrics data.
2 Injects the OpenShift service CA for configuring the TLS in the Prometheus receiver.
3 Configures the Prometheus receiver to scrape the federate endpoint from the in-cluster monitoring stack.
4 Uses the Prometheus query language to select the metrics to be scraped. See the in-cluster monitoring documentation for more details and limitations of the federate endpoint.
5 Configures the debug exporter to print the metrics to the standard output.

Additional resources