[management,signal,proxy] add pyroscope profiling (#7536)

proxy/internal/metrics/client_metrics_test.go

@@ -0,0 +1,49 @@
+package metrics_test
+
+import (
+	"context"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	sdkmetric "go.opentelemetry.io/otel/sdk/metric"
+	"go.opentelemetry.io/otel/sdk/metric/metricdata"
+
+	"github.com/netbirdio/netbird/proxy/internal/metrics"
+)
+
+func TestRegisterClientObserver(t *testing.T) {
+	reader := sdkmetric.NewManualReader()
+	provider := sdkmetric.NewMeterProvider(sdkmetric.WithReader(reader))
+	m, err := metrics.New(context.Background(), provider.Meter("test"))
+	require.NoError(t, err)
+
+	clients := 2
+	require.NoError(t, m.RegisterClientObserver(func() int { return clients }))
+
+	var rm metricdata.ResourceMetrics
+	require.NoError(t, reader.Collect(context.Background(), &rm))
+	assert.Equal(t, int64(2), gaugeValue(t, rm, "proxy.clients.count"), "gauge must report the current client count")
+
+	clients = 1
+	require.NoError(t, reader.Collect(context.Background(), &rm))
+	assert.Equal(t, int64(1), gaugeValue(t, rm, "proxy.clients.count"), "gauge must follow the client count on the next collection")
+}
+
+func gaugeValue(t *testing.T, rm metricdata.ResourceMetrics, name string) int64 {
+	t.Helper()
+
+	for _, sm := range rm.ScopeMetrics {
+		for _, mtr := range sm.Metrics {
+			if mtr.Name != name {
+				continue
+			}
+			gauge, ok := mtr.Data.(metricdata.Gauge[int64])
+			require.True(t, ok, "%s must be an int64 gauge", name)
+			require.Len(t, gauge.DataPoints, 1, "%s must have a single data point", name)
+			return gauge.DataPoints[0].Value
+		}
+	}
+	t.Fatalf("gauge %s not found", name)
+	return 0
+}

proxy/internal/metrics/metrics.go

@@ -196,6 +196,21 @@ func (m *Metrics) RecordAddPeerDuration(d time.Duration, err error) {
 	))
 }
 
+// RegisterClientObserver reports the number of embedded clients as a gauge.
+// clientCount runs on every collection cycle, so it must stay cheap.
+func (m *Metrics) RegisterClientObserver(clientCount func() int) error {
+	_, err := m.meter.Int64ObservableGauge(
+		"proxy.clients.count",
+		metric.WithUnit("1"),
+		metric.WithDescription("Current number of embedded NetBird clients running on the netbird proxy"),
+		metric.WithInt64Callback(func(_ context.Context, o metric.Int64Observer) error {
+			o.Observe(int64(clientCount()))
+			return nil
+		}),
+	)
+	return err
+}
+
 func (m *Metrics) initL4Metrics(meter metric.Meter) error {
 	var err error