Interface Metrics

cmd/frontend/internal/env/config.go

@@ -44,4 +44,6 @@ type Config struct {
 	DelayConnectivity     time.Duration `default:"1s" desc:"Delay between checks with connectivity"`
 	DelayNoConnectivity   time.Duration `default:"3s" desc:"Delay between checks without connectivity"`
 	MaxSessionErrors      int           `default:"5" desc:"Max session errors when checking Bird until denounce"`
+	MetricsEnabled        bool          `default:"false" desc:"Enable the metrics collection" split_words:"true"`
+	MetricsPort           int           `default:"2224" desc:"Specify the port used to expose the metrics" split_words:"true"`
 }

cmd/frontend/main.go

@@ -28,6 +28,8 @@ import (
 
 	"github.com/go-logr/logr"
 	"github.com/kelseyhightower/envconfig"
+	"go.opentelemetry.io/otel/attribute"
+	"go.opentelemetry.io/otel/metric"
 	"google.golang.org/grpc"
 	"google.golang.org/grpc/backoff"
 	"google.golang.org/grpc/keepalive"
@@ -38,7 +40,9 @@ import (
 	"github.com/nordix/meridio/cmd/frontend/internal/frontend"
 	"github.com/nordix/meridio/pkg/health"
 	"github.com/nordix/meridio/pkg/health/connection"
+	linuxKernel "github.com/nordix/meridio/pkg/kernel"
 	"github.com/nordix/meridio/pkg/log"
+	"github.com/nordix/meridio/pkg/metrics"
 	"github.com/nordix/meridio/pkg/retry"
 	"github.com/nordix/meridio/pkg/security/credentials"
 )
@@ -85,6 +89,11 @@ func main() {
 	)
 	defer cancel()
 
+	hostname, err := os.Hostname()
+	if err != nil {
+		log.Fatal(logger, "Unable to get hostname", "error", err)
+	}
+
 	// create and start health server
 	ctx = health.CreateChecker(ctx)
 	if err := health.RegisterReadinesSubservices(ctx, health.FEReadinessServices...); err != nil {
@@ -155,6 +164,43 @@ func main() {
 	// start watching events of interest via NSP
 	go watchConfig(ctx, cancel, c, fe)
 
+	interfaceMetrics := linuxKernel.NewInterfaceMetrics([]metric.ObserveOption{
+		metric.WithAttributes(attribute.String("Hostname", hostname)),
+		metric.WithAttributes(attribute.String("Trench", config.TrenchName)),
+		metric.WithAttributes(attribute.String("Attractor", config.AttractorName)),
+	})
+	interfaceMetrics.Register(config.ExternalInterface)
+
+	if config.MetricsEnabled {
+		func() {
+			_, err = metrics.Init(ctx)
+			if err != nil {
+				logger.Error(err, "Unable to init metrics collector")
+				cancel()
+				return
+			}
+
+			err = interfaceMetrics.Collect()
+			if err != nil {
+				logger.Error(err, "Unable to start interface metrics collector")
+				cancel()
+				return
+			}
+
+			metricsServer := metrics.Server{
+				IP:   "",
+				Port: config.MetricsPort,
+			}
+			go func() {
+				err := metricsServer.Start(ctx)
+				if err != nil {
+					logger.Error(err, "Unable to start metrics server")
+					cancel()
+				}
+			}()
+		}()
+	}
+
 	<-ctx.Done()
 	logger.Info("FE shutting down")
 }

cmd/stateless-lb/main.go

@@ -56,10 +56,13 @@ import (
 	"github.com/nordix/meridio/pkg/networking"
 	"github.com/nordix/meridio/pkg/nsm"
 	"github.com/nordix/meridio/pkg/nsm/interfacemonitor"
+	nsmmetrics "github.com/nordix/meridio/pkg/nsm/metrics"
 	"github.com/nordix/meridio/pkg/retry"
 	"github.com/nordix/meridio/pkg/security/credentials"
 	"github.com/sirupsen/logrus"
 	"github.com/vishvananda/netlink"
+	"go.opentelemetry.io/otel/attribute"
+	"go.opentelemetry.io/otel/metric"
 	"google.golang.org/grpc"
 	"google.golang.org/grpc/backoff"
 	"google.golang.org/grpc/keepalive"
@@ -177,6 +180,12 @@ func main() {
 		log.Fatal(logger, "Unable to init lb target metrics", "error", err)
 	}
 
+	interfaceMetrics := linuxKernel.NewInterfaceMetrics([]metric.ObserveOption{
+		metric.WithAttributes(attribute.String("Hostname", hostname)),
+		metric.WithAttributes(attribute.String("Trench", config.TrenchName)),
+		metric.WithAttributes(attribute.String("Conduit", config.ConduitName)),
+	})
+
 	lbFactory := nfqlb.NewLbFactory(nfqlb.WithNFQueue(config.Nfqueue))
 	nfa, err := nfqlb.NewNetfilterAdaptor(nfqlb.WithNFQueue(config.Nfqueue), nfqlb.WithNFQueueFanout(config.NfqueueFanout))
 	if err != nil {
@@ -208,6 +217,7 @@ func main() {
 			noop.MECHANISM:       null.NewServer(),
 		}),
 		interfaceMonitorEndpoint,
+		nsmmetrics.NewServer(interfaceMetrics),
 		sendfd.NewServer(),
 	}
 
@@ -295,6 +305,13 @@ func main() {
 				return
 			}
 
+			err = interfaceMetrics.Collect()
+			if err != nil {
+				logger.Error(err, "Unable to start interface metrics collector")
+				cancel()
+				return
+			}
+
 			metricsServer := metrics.Server{
 				IP:   "",
 				Port: config.MetricsPort,

config/templates/charts/meridio/deployment/stateless-lb-frontend.yaml

@@ -184,6 +184,9 @@ spec:
         - name: frontend
           image: {{ .Values.registry }}/{{ .Values.repository }}/{{ .Values.frontEnd.image }}:{{ .Values.version }}
           imagePullPolicy:  # Kubernetes default according to image tag
+          ports:
+          - name: metrics
+            containerPort: 2224
           startupProbe:  # will be filled by operator if not specified
             exec:
               command:
@@ -244,6 +247,8 @@ spec:
               value:  # to be filled by operator
             - name: NFE_LOG_LEVEL
               value:  # to be filled by operator
+            - name: NFE_METRICS_ENABLED
+              value: "true"
           securityContext:
             runAsNonRoot: true
             readOnlyRootFilesystem: true