Add additional labels from BMC and Server resources to metrics

cmd/main.go

@@ -83,6 +83,8 @@ func main() { // nolint: gocyclo
 		eventPort                          int
 		eventURL                           string
 		eventProtocol                      string
+		redfishMetricLabelsFromBMC         string
+		redfishMetricLabelsFromServer      string
 		registryClientTimeout              time.Duration
 		registryDataMaxAge                 time.Duration
 		registryResyncInterval             time.Duration
@@ -143,6 +145,16 @@ func main() { // nolint: gocyclo
 	flag.IntVar(&eventPort, "event-port", 10001, "The port to use for the server events endpoint for alerts and metrics.")
 	flag.StringVar(&eventProtocol, "event-protocol", "http",
 		"The protocol to use for the server events endpoint for alerts and metrics.")
+	flag.StringVar(&redfishMetricLabelsFromBMC, "redfish-metric-labels-from-bmc", "",
+		"Comma-separated list of 'kubernetes-label-key=prometheus-label-name' pairs. "+
+			"Each pair adds an additional label dimension to Redfish telemetry metrics, "+
+			"sourced from the matching label on the BMC resource. "+
+			"Example: topology.kubernetes.io/region=region,topology.kubernetes.io/zone=zone")
+	flag.StringVar(&redfishMetricLabelsFromServer, "redfish-metric-labels-from-server", "",
+		"Comma-separated list of 'kubernetes-label-key=prometheus-label-name' pairs. "+
+			"Each pair adds an additional label dimension to Redfish telemetry metrics, "+
+			"sourced from the matching label on the Server resource linked via spec.bmcRef.name. "+
+			"Example: metadata.metal.ironcore.dev/location=location,metadata.metal.ironcore.dev/rack=rack")
 	flag.StringVar(&probeImage, "probe-image", "", "Image for the first boot probing of a Server.")
 	flag.StringVar(&probeOSImage, "probe-os-image", "", "OS image for the first boot probing of a Server.")
 	flag.StringVar(&managerNamespace, "manager-namespace", "default", "Namespace the manager is running in.")
@@ -708,9 +720,21 @@ func main() { // nolint: gocyclo
 	}
 
 	if eventURL != "" {
+		bmcLabelMappings, err := serverevents.ParseLabelMappings(redfishMetricLabelsFromBMC)
+		if err != nil {
+			setupLog.Error(err, "Invalid --redfish-metric-labels-from-bmc")
+			os.Exit(1)
+		}
+		serverLabelMappings, err := serverevents.ParseLabelMappings(redfishMetricLabelsFromServer)
+		if err != nil {
+			setupLog.Error(err, "Invalid --redfish-metric-labels-from-server")
+			os.Exit(1)
+		}
 		if err := mgr.Add(manager.RunnableFunc(func(ctx context.Context) error {
 			setupLog.Info("starting event server for alerts and metrics", "EventURL", eventURL)
-			eventServer := serverevents.NewServer(setupLog, fmt.Sprintf(":%d", eventPort))
+			eventServer := serverevents.NewServer(
+				setupLog, fmt.Sprintf(":%d", eventPort), mgr.GetClient(), bmcLabelMappings, serverLabelMappings,
+			)
 			if err := eventServer.Start(ctx); err != nil {
 				return fmt.Errorf("unable to start event server: %w", err)
 			}

dist/chart/templates/_helpers.tpl

@@ -48,3 +48,17 @@ app.kubernetes.io/instance: {{ .Release.Name }}
     $hasValidating = true }}{{- end }}
 {{- end }}
 {{ $hasValidating }}}}{{- end }}
+
+{{/*
+chart.redfishLabelFlag renders a single CLI flag string from a map of
+kubernetes-label-key -> prometheus-label-name entries.
+
+Usage: {{ include "chart.redfishLabelFlag" (dict "flag" "redfish-metric-labels-from-bmc" "map" .Values.redfishLabels.bmc) }}
+*/}}
+{{- define "chart.redfishLabelFlag" -}}
+{{- $pairs := list -}}
+{{- range $k, $v := .map -}}
+  {{- $pairs = append $pairs (printf "%s=%s" $k $v) -}}
+{{- end -}}
+{{- printf "--%s=%s" .flag (join "," ($pairs | sortAlpha)) -}}
+{{- end }}

dist/chart/templates/manager/manager.yaml

@@ -34,6 +34,12 @@ spec:
         {{- range .Values.controllerManager.manager.args }}
         - {{ . }}
         {{- end }}
+        {{- if .Values.redfishLabels.bmc }}
+        - {{ include "chart.redfishLabelFlag" (dict "flag" "redfish-metric-labels-from-bmc" "map" .Values.redfishLabels.bmc) | quote }}
+        {{- end }}
+        {{- if .Values.redfishLabels.server }}
+        - {{ include "chart.redfishLabelFlag" (dict "flag" "redfish-metric-labels-from-server" "map" .Values.redfishLabels.server) | quote }}
+        {{- end }}
         command:
         - /manager
         image: {{ .Values.controllerManager.manager.image.repository }}:{{ .Values.controllerManager.manager.image.tag }}

dist/chart/values.yaml

@@ -67,6 +67,23 @@ crd:
   # (Certificates, Issuers, ...) due to garbage collection.
   keep: true
 
+# [REDFISH LABEL ENRICHMENT]: Optional label enrichment for Redfish telemetry metrics.
+# Define mappings from Kubernetes resource label keys to Prometheus label names.
+# Mapped labels are appended as additional dimensions to redfish_monitor_reading and
+# redfish_event_alert_total metrics. Leave empty ({}) to disable enrichment.
+redfishLabels:
+  # Labels sourced from the BMC resource (matched by hostname == BMC resource name).
+  bmc: {}
+  # Example:
+  #   topology.kubernetes.io/region: region
+  #   topology.kubernetes.io/zone: zone
+
+  # Labels sourced from the Server resource linked via spec.bmcRef.name.
+  server: {}
+  # Example:
+  #   metadata.metal.ironcore.dev/location: location
+  #   metadata.metal.ironcore.dev/rack: rack
+
 # [METRICS]: Set to true to generate manifests for exporting metrics.
 # To disable metrics export set false, and ensure that the
 # ControllerManager argument "--metrics-bind-address=:8443" is removed.
@@ -101,4 +118,3 @@ ignition:
   # Template content that can be customized - this will be created as a ConfigMap
   # and mounted to override the default template
   # template: |
-

docs/observability/metrics.md

@@ -281,6 +281,97 @@ rate(metal_server_reconciliation_total{result=~"error_.*"}[5m])
 count(metal_server_state{state="Available"} == 1)
 ```
 
+## Redfish Telemetry Metrics
+
+When the operator is configured with an event URL (`--event-url`), it subscribes to Redfish MetricReport and Alert events from each BMC and exposes two additional metrics.
+
+### Sensor Readings (`redfish_monitor_reading`)
+
+**Type:** Gauge
+**Description:** Latest sensor value pushed via a Redfish MetricReport event.
+**Fixed labels:**
+- `hostname`: BMC Kubernetes resource name
+- `metric_id`: Redfish metric ID (e.g., `CPU1Temp`)
+- `type`: Metric type (e.g., `Temperature`, `Voltage`)
+- `unit`: Unit of measure (e.g., `Cel`, `V`)
+- `origin_context`: Originating hardware component path
+
+**Dynamic labels:** Additional label dimensions can be injected from the BMC or Server resource (see [Label Enrichment](#label-enrichment) below).
+
+**Example values:**
+```text
+redfish_monitor_reading{hostname="node001-bmc", metric_id="CPU1Temp", type="Temperature", unit="Cel", origin_context="/Chassis/1/Thermal"} 42.5
+redfish_monitor_reading{hostname="node001-bmc", metric_id="FanSpeed1", type="Rotational", unit="RPM", origin_context="/Chassis/1/Thermal"} 3200
+```
+
+**Use cases:**
+- Alert on thermal readings exceeding thresholds: `redfish_monitor_reading{type="Temperature"} > 80`
+- Track fan speeds: `redfish_monitor_reading{type="Rotational"}`
+- Compare readings across regions or racks when enriched with topology labels
+
+### Alert Event Counter (`redfish_event_alert_total`)
+
+**Type:** Counter
+**Description:** Total count of Redfish alert/event messages received from each BMC.
+**Fixed labels:**
+- `hostname`: BMC Kubernetes resource name
+- `severity`: Event severity (e.g., `OK`, `Warning`, `Critical`)
+- `message_id`: Redfish MessageId (e.g., `Alert.1.0.ResourceStatusChangedOK`)
+- `component`: Originating hardware component
+
+**Dynamic labels:** Same enrichment as `redfish_monitor_reading`.
+
+**Example values:**
+```text
+redfish_event_alert_total{hostname="node001-bmc", severity="Warning", message_id="ThermalEvents.1.0.TemperatureAboveUpperCautionThreshold", component="/Chassis/1/Thermal/CPU1Temp"} 3
+redfish_event_alert_total{hostname="node001-bmc", severity="OK", message_id="Alert.1.0.ResourceStatusChangedOK", component="/Systems/1"} 12
+```
+
+**Use cases:**
+- Alert on sustained critical events: `increase(redfish_event_alert_total{severity="Critical"}[5m]) > 0`
+- Track warning frequency per host: `rate(redfish_event_alert_total{severity="Warning"}[1h])`
+
+### Label Enrichment
+
+When managing a large number of servers, it is often necessary to filter dashboard panels and alert rules by topology or location (e.g., region, availability zone, rack). Both Redfish metrics support optional dynamic label dimensions sourced from Kubernetes resources for exactly this purpose — enabling operators to slice telemetry by any organisational dimension without modifying the operator itself.
+
+This is configured via two CLI flags:
+
+| Flag | Source resource | Match key |
+|------|----------------|-----------|
+| `--redfish-metric-labels-from-bmc` | `BMC` resource | resource name == `hostname` label |
+| `--redfish-metric-labels-from-server` | `Server` resource | `spec.bmcRef.name` == `hostname` label |
+
+**Flag format:** `kubernetes-label-key=prometheus-label-name,...`
+
+**Example:**
+```bash
+--redfish-metric-labels-from-bmc=topology.kubernetes.io/region=region,topology.kubernetes.io/zone=zone
+--redfish-metric-labels-from-server=metadata.metal.ironcore.dev/location=location,metadata.metal.ironcore.dev/rack=rack
+```
+
+When configured, every Redfish metric gains the extra label columns. If a label key is missing from the resource, the value is emitted as an empty string — missing labels never block metric emission.
+
+Labels are read from the controller-runtime informer cache, which is watch-based and always reflects the current cluster state. There is no TTL — label changes on BMC or Server resources are visible immediately.
+
+#### Helm chart configuration
+
+```yaml
+redfishLabels:
+  bmc:
+    topology.kubernetes.io/region: region
+    topology.kubernetes.io/zone: zone
+  server:
+    metadata.metal.ironcore.dev/location: location
+    metadata.metal.ironcore.dev/rack: rack
+```
+
+#### Example enriched output
+
+```text
+redfish_monitor_reading{hostname="node001-bmc", metric_id="CPU1Temp", type="Temperature", unit="Cel", origin_context="/Chassis/1/Thermal", region="eu-de-1", zone="eu-de-1a", location="building-b", rack="row3-rack7"} 42.5
+```
+
 ## Implementation Details
 
 ### Metric Collection Strategy