feat: ping time metric

pkg/hive/hive.go

@@ -284,12 +284,16 @@ func (s *Service) checkAndAddPeers(ctx context.Context, peers pb.Peers) {
 			ctx, cancel := context.WithTimeout(ctx, pingTimeout)
 			defer cancel()
 
+			start := time.Now()
+
 			// check if the underlay is usable by doing a raw ping using libp2p
 			if _, err = s.streamer.Ping(ctx, multiUnderlay); err != nil {
+				s.metrics.PingFailureTime.Observe(time.Since(start).Seconds())
 				s.metrics.UnreachablePeers.Inc()
 				s.logger.Debugf("hive: peer %s: underlay %s not reachable", hex.EncodeToString(newPeer.Overlay), multiUnderlay)
 				return
 			}
+			s.metrics.PingTime.Observe(time.Since(start).Seconds())
 
 			bzzAddress := bzz.Address{
 				Overlay:     swarm.NewAddress(newPeer.Overlay),

pkg/hive/metrics.go

@@ -17,6 +17,9 @@ type metrics struct {
 	PeersHandler      prometheus.Counter
 	PeersHandlerPeers prometheus.Counter
 	UnreachablePeers  prometheus.Counter
+
+	PingTime        prometheus.Histogram
+	PingFailureTime prometheus.Histogram
 }
 
 func newMetrics() metrics {
@@ -59,6 +62,18 @@ func newMetrics() metrics {
 			Name:      "unreachable_peers_count",
 			Help:      "Number of peers that are unreachable.",
 		}),
+		PingTime: prometheus.NewHistogram(prometheus.HistogramOpts{
+			Namespace: m.Namespace,
+			Subsystem: subsystem,
+			Name:      "ping_time",
+			Help:      "The time spent for pings.",
+		}),
+		PingFailureTime: prometheus.NewHistogram(prometheus.HistogramOpts{
+			Namespace: m.Namespace,
+			Subsystem: subsystem,
+			Name:      "fail_ping_time",
+			Help:      "The time spent for unsuccessful pings.",
+		}),
 	}
 }