go/p2p: collect Prometheus metrics

.changelog/5327.feature.md

@@ -0,0 +1,13 @@
+go/p2p: implement various metrics collection
+
+The following Prometheus metrics were added:
+
+- `oasis_p2p_peers`
+- `oasis_p2p_blocked_peers`
+- `oasis_p2p_connections`
+- `oasis_p2p_topics`
+- `oasis_p2p_protocols`
+
+See [metrics documentation] for descriptions of metrics.
+
+[metrics documentation]: https://docs.oasis.io/core/oasis-node/metrics

docs/oasis-node/metrics.md

@@ -66,6 +66,11 @@ oasis_node_net_receive_bytes_total | Gauge | Received data for each network devi
 oasis_node_net_receive_packets_total | Gauge | Received data for each network device as reported by /proc/net/dev (packets). | device | [oasis-node/cmd/common/metrics](https://github.com/oasisprotocol/oasis-core/tree/master/go/oasis-node/cmd/common/metrics/net.go)
 oasis_node_net_transmit_bytes_total | Gauge | Transmitted data for each network device as reported by /proc/net/dev (bytes). | device | [oasis-node/cmd/common/metrics](https://github.com/oasisprotocol/oasis-core/tree/master/go/oasis-node/cmd/common/metrics/net.go)
 oasis_node_net_transmit_packets_total | Gauge | Transmitted data for each network device as reported by /proc/net/dev (packets). | device | [oasis-node/cmd/common/metrics](https://github.com/oasisprotocol/oasis-core/tree/master/go/oasis-node/cmd/common/metrics/net.go)
+oasis_p2p_blocked_peers | Gauge | Number of blocked P2P peers. |  | [p2p](https://github.com/oasisprotocol/oasis-core/tree/master/go/p2p/metrics.go)
+oasis_p2p_connections | Gauge | Number of P2P connections. |  | [p2p](https://github.com/oasisprotocol/oasis-core/tree/master/go/p2p/metrics.go)
+oasis_p2p_peers | Gauge | Number of connected P2P peers. |  | [p2p](https://github.com/oasisprotocol/oasis-core/tree/master/go/p2p/metrics.go)
+oasis_p2p_protocols | Gauge | Number of supported P2P protocols. |  | [p2p](https://github.com/oasisprotocol/oasis-core/tree/master/go/p2p/metrics.go)
+oasis_p2p_topics | Gauge | Number of supported P2P topics. |  | [p2p](https://github.com/oasisprotocol/oasis-core/tree/master/go/p2p/metrics.go)
 oasis_registry_entities | Gauge | Number of registry entities. |  | [registry](https://github.com/oasisprotocol/oasis-core/tree/master/go/registry/metrics.go)
 oasis_registry_nodes | Gauge | Number of registry nodes. |  | [registry](https://github.com/oasisprotocol/oasis-core/tree/master/go/registry/metrics.go)
 oasis_registry_runtimes | Gauge | Number of registry runtimes. |  | [registry](https://github.com/oasisprotocol/oasis-core/tree/master/go/registry/metrics.go)

go/p2p/metrics.go

@@ -0,0 +1,78 @@
+package p2p
+
+import (
+	"sync"
+	"time"
+
+	"github.com/prometheus/client_golang/prometheus"
+
+	cmmetrics "github.com/oasisprotocol/oasis-core/go/oasis-node/cmd/common/metrics"
+)
+
+const metricsUpdateInterval = 60 * time.Second
+
+var (
+	peersMetric = prometheus.NewGauge(prometheus.GaugeOpts{
+		Name: "oasis_p2p_peers",
+		Help: "Number of connected P2P peers.",
+	})
+	blockedPeersMetric = prometheus.NewGauge(prometheus.GaugeOpts{
+		Name: "oasis_p2p_blocked_peers",
+		Help: "Number of blocked P2P peers.",
+	})
+	connectionsMetric = prometheus.NewGauge(prometheus.GaugeOpts{
+		Name: "oasis_p2p_connections",
+		Help: "Number of P2P connections.",
+	})
+	topicsMetric = prometheus.NewGauge(prometheus.GaugeOpts{
+		Name: "oasis_p2p_topics",
+		Help: "Number of supported P2P topics.",
+	})
+	protocolsMetric = prometheus.NewGauge(prometheus.GaugeOpts{
+		Name: "oasis_p2p_protocols",
+		Help: "Number of supported P2P protocols.",
+	})
+
+	p2pCollectors = []prometheus.Collector{
+		peersMetric,
+		blockedPeersMetric,
+		connectionsMetric,
+		topicsMetric,
+		protocolsMetric,
+	}
+
+	metricsOnce sync.Once
+)
+
+func (p *p2p) metricsWorker() {
+	defer close(p.metricsClosedCh)
+	if !cmmetrics.Enabled() {
+		return
+	}
+
+	metricsOnce.Do(func() {
+		prometheus.MustRegister(p2pCollectors...)
+	})
+	// Update the metrics once on initialization.
+	p.updateMetrics()
+
+	t := time.NewTicker(metricsUpdateInterval)
+	defer t.Stop()
+
+	for {
+		select {
+		case <-p.ctx.Done():
+			return
+		case <-t.C:
+			p.updateMetrics()
+		}
+	}
+}
+
+func (p *p2p) updateMetrics() {
+	peersMetric.Set(float64(len(p.host.Network().Peers())))
+	blockedPeersMetric.Set(float64(len(p.gater.ListBlockedPeers())))
+	connectionsMetric.Set(float64(len(p.host.Network().Conns())))
+	topicsMetric.Set(float64(len(p.peerMgr.Topics())))
+	protocolsMetric.Set(float64(len(p.peerMgr.Protocols())))
+}

go/p2p/p2p.go

@@ -67,9 +67,10 @@ func DebugForceAllowUnroutableAddresses() {
 type p2p struct {
 	sync.RWMutex
 
-	ctx       context.Context
-	ctxCancel context.CancelFunc
-	quitCh    chan struct{}
+	ctx             context.Context
+	ctxCancel       context.CancelFunc
+	quitCh          chan struct{}
+	metricsClosedCh chan struct{}
 
 	chainContext string
 	signer       signature.Signer
@@ -111,7 +112,7 @@ func (p *p2p) Stop() {
 
 	var wg sync.WaitGroup
 	defer wg.Wait()
-	wg.Add(2)
+	wg.Add(3)
 
 	go func() {
 		defer wg.Done()
@@ -122,6 +123,11 @@ func (p *p2p) Stop() {
 		defer wg.Done()
 		_ = p.host.Close() // This blocks until the host stops.
 	}()
+
+	go func() {
+		defer wg.Done()
+		<-p.metricsClosedCh
+	}()
 }
 
 // Implements api.Service.
@@ -414,6 +420,7 @@ func New(identity *identity.Identity, consensus consensus.Backend, store *persis
 		ctx:               ctx,
 		ctxCancel:         ctxCancel,
 		quitCh:            make(chan struct{}),
+		metricsClosedCh:   make(chan struct{}),
 		chainContext:      chainContext,
 		signer:            identity.P2PSigner,
 		host:              host,
@@ -435,6 +442,8 @@ func New(identity *identity.Identity, consensus consensus.Backend, store *persis
 		)
 	}
 
+	go p.metricsWorker()
+
 	return p, nil
 }