Monitoring

The nbe-operator exposes Prometheus metrics and health check endpoints for observability.

Operator Metrics

The operator exposes metrics at :8080/metrics in Prometheus format.

Prometheus Annotations

When metrics.enabled: true (the default), Prometheus scrape annotations are added to the operator pod:

prometheus.io/scrape: "true"
prometheus.io/port: "8080"
prometheus.io/path: "/metrics"

Key	Type	Default	Description
metrics.enabled	bool	true	Enable Prometheus annotations on the operator pod
metrics.podAnnotations	bool	true	Add standard prometheus.io/* annotations

ServiceMonitor

For clusters using the Prometheus Operator, create a ServiceMonitor:

serviceMonitor:
  enabled: true
  interval: "30s"
  scrapeTimeout: "10s"

Full ServiceMonitor configuration:

Key	Type	Default	Description
serviceMonitor.enabled	bool	false	Create a ServiceMonitor resource
serviceMonitor.namespace	string	Release namespace	Target namespace
serviceMonitor.labels	object	{}	Labels for ServiceMonitor selection
serviceMonitor.interval	string	30s	Scrape interval
serviceMonitor.scrapeTimeout	string	10s	Scrape timeout
serviceMonitor.scheme	string	http	HTTP scheme
serviceMonitor.honorLabels	bool	true	Honor labels from metrics

NetBox Application Metrics

Enable the NetBox /metrics endpoint separately via the NetBoxEnterprise spec:

netboxEnterprise:
  spec:
    netbox:
      config:
        metricsEnabled: true

The operator aggregates metrics from NetBox and Diode deployments and exposes them at its own /metrics endpoint, so Prometheus only needs to scrape the operator.

note

Deployment metrics aggregation (collecting metrics from NetBox/Diode pods) is a new feature. Contact NetBox Labs support for early access.

Diode Component Metrics

Enable per-component metrics with telemetry configuration:

netboxEnterprise:
  spec:
    diode:
      config:
        ingester:
          telemetryConfig:
            metricsEnabled: true
        reconciler:
          telemetryConfig:
            metricsEnabled: true
        auth:
          telemetryConfig:
            metricsEnabled: true

Supported exporters: prometheus, otlp, console, none.

Health Check Endpoints

The operator exposes two health endpoints:

Endpoint	Port	Purpose
/healthz	8081	Liveness probe — is the operator process alive?
/readyz	8081	Readiness probe — is the operator ready to serve?

These are configured as Kubernetes liveness and readiness probes on the operator pod.

Operator Log Levels

Adjust operator verbosity for debugging:

operator:
  logging:
    level: "debug"     # Or: info, info,kube=warn, operator=debug,info
    format: "json"     # Or: auto, compact, pretty, gcp, aws, otlp

Format	Use Case
auto	Detects environment (JSON in Kubernetes, compact locally)
json	Structured logging for log aggregation (Elasticsearch, Loki)
compact	Single-line human-readable for local development
pretty	Multi-line verbose for debugging
gcp	Google Cloud Logging format (auto-ingested in GKE)
aws	CloudWatch-optimized JSON (auto-ingested in EKS)
otlp	OpenTelemetry export for Azure Monitor, Jaeger, etc.

Next Steps

• Troubleshooting — Using metrics and logs to diagnose issues
• Helm Values Reference — Full operator configuration