[DRAFT] feat(edge): strengthen edge reporter with metrics, backoff, structured logging

pkg/edge/config.go

@@ -21,6 +21,26 @@ type Config struct {
 
 	// Cloud holds the upstream (L2/L3) connection settings.
 	Cloud CloudConfig `json:"cloud"`
+
+	// StatusInterval is how often status snapshots are sent to the Fleet Manager.
+	// Defaults to 5 minutes if zero.
+	StatusInterval time.Duration `json:"status_interval"`
+
+	// MaxRetries is the maximum number of delivery retries for failed HTTP calls.
+	// Defaults to 3 if zero.
+	MaxRetries int `json:"max_retries"`
+
+	// RetryBaseDelay is the initial backoff duration for retries.
+	// Each retry doubles this with jitter. Defaults to 1s if zero.
+	RetryBaseDelay time.Duration `json:"retry_base_delay"`
+
+	// EventQueueSize is the capacity of the buffered event queue for async reporting.
+	// Defaults to 256 if zero.
+	EventQueueSize int `json:"event_queue_size"`
+
+	// EventWorkers is the number of goroutines draining the async event queue.
+	// Defaults to 1 if zero.
+	EventWorkers int `json:"event_workers"`
 }
 
 // CloudConfig holds the upstream cloud/gateway connection settings.
@@ -46,5 +66,10 @@ func DefaultConfig() Config {
 			Endpoint:          "http://localhost:8080",
 			HeartbeatInterval: 30 * time.Second,
 		},
+		StatusInterval:  5 * time.Minute,
+		MaxRetries:      3,
+		RetryBaseDelay:  1 * time.Second,
+		EventQueueSize:  256,
+		EventWorkers:    1,
 	}
-}
+}

pkg/edge/metrics.go

@@ -0,0 +1,72 @@
+package edge
+
+import (
+	"expvar"
+	"sync"
+	"sync/atomic"
+	"time"
+)
+
+// ReporterMetrics holds live counters and gauges for the edge reporter.
+// Exported via expvar under "edgereporter" so operators can scrape them
+// alongside any Prometheus /debug/vars exporter.
+type ReporterMetrics struct {
+	mu sync.RWMutex
+
+	// Counters (cumulative)
+	RegistrationsTotal  atomic.Int64 `json:"registrations_total"`
+	HeartbeatsTotal     atomic.Int64 `json:"heartbeats_total"`
+	StatusReportsTotal  atomic.Int64 `json:"status_reports_total"`
+	EventsTotal         atomic.Int64 `json:"events_total"`
+	QueuedEventsTotal   atomic.Int64 `json:"queued_events_total"`
+	DeliveriesTotal     atomic.Int64 `json:"deliveries_total"`      // all HTTP POST attempts
+	DeliveriesOK        atomic.Int64 `json:"deliveries_ok"`         // 2xx responses
+	DeliveriesFailed    atomic.Int64 `json:"deliveries_failed"`     // non-2xx or network errors
+	RetriesTotal        atomic.Int64 `json:"retries_total"`         // retry attempts due to failure
+	DroppedEventsTotal  atomic.Int64 `json:"dropped_events_total"`  // events dropped when queue full
+
+	// Gauges (snapshots)
+	QueueDepth     atomic.Int64 `json:"queue_depth"`
+	LastDeliveryOK atomic.Bool  `json:"last_delivery_ok"`
+
+	// Latency buckets (ms) for the most recent deliveries (ring buffer)
+	latencyMu   sync.Mutex
+	latencyRing [64]time.Duration
+	latencyIdx  int
+}
+
+var metrics = &ReporterMetrics{}
+
+func init() {
+	expvar.Publish("edgereporter", expvar.Func(func() any {
+		return metrics.Snapshot()
+	}))
+}
+
+// Snapshot returns a consistent read of all metrics.
+func (m *ReporterMetrics) Snapshot() map[string]any {
+	m.mu.RLock()
+	defer m.mu.RUnlock()
+
+	return map[string]any{
+		"registrations_total":  m.RegistrationsTotal.Load(),
+		"heartbeats_total":     m.HeartbeatsTotal.Load(),
+		"status_reports_total": m.StatusReportsTotal.Load(),
+		"events_total":         m.EventsTotal.Load(),
+		"queued_events_total":  m.QueuedEventsTotal.Load(),
+		"deliveries_total":     m.DeliveriesTotal.Load(),
+		"deliveries_ok":        m.DeliveriesOK.Load(),
+		"deliveries_failed":    m.DeliveriesFailed.Load(),
+		"retries_total":        m.RetriesTotal.Load(),
+		"dropped_events_total": m.DroppedEventsTotal.Load(),
+		"queue_depth":          m.QueueDepth.Load(),
+		"last_delivery_ok":     m.LastDeliveryOK.Load(),
+	}
+}
+
+func (m *ReporterMetrics) recordLatency(d time.Duration) {
+	m.latencyMu.Lock()
+	m.latencyRing[m.latencyIdx%64] = d
+	m.latencyIdx++
+	m.latencyMu.Unlock()
+}