EL: add some discovery metrics to check disv4/5 performance

execution_chain/networking/discoveryv4/kademlia.nim

@@ -41,6 +41,7 @@ type
   Node* = ref object
     node*: ENode
     id*: NodeId
+    fromDiscv5*: bool
 
   RoutingTable = object
     thisNode: Node

execution_chain/networking/eth1_discovery.nim

@@ -47,6 +47,10 @@ type
     discv5: DiscV5
     compatibleForkId: CompatibleForkIdProc
 
+declareCounter disc_incoming_packets, "Number of incoming packets on discovery port", labels = ["type"]
+declareCounter discv5_nodes_discovered, "Number of nodes discovered via discv5", labels = ["type"]
+declareCounter discv4_nodes_discovered, "Number of nodes discovered via discv4"
+
 #------------------------------------------------------------------------------
 # Private functions
 #------------------------------------------------------------------------------
@@ -65,6 +69,7 @@ func to(node: NodeV5, _: type NodeV4): ENodeResult[NodeV4] =
   let v4 = NodeV4(
     id: node.id,
     node: ?ENode.fromEnr(node.record),
+    fromDiscv5: true,
   )
   ok(v4)
 
@@ -90,8 +95,15 @@ proc processClient(
   if discv4.isErr:
     # unhandled buf will be handled by discv5
     let addrv5 = raddr.to(AddressV5)
-    proto.discv5.receiveV5(addrv5, buf).isOkOr:
-      debug "Discovery receive error", discv4=discv4.error, discv5=error
+    let res = proto.discv5.receiveV5(addrv5, buf)
+    if res.isErr:
+      debug "Discovery receive error", discv4=discv4.error, discv5=res.error, address=addrv5
+      disc_incoming_packets.inc(labelValues = ["spam"])
+    else:
+      disc_incoming_packets.inc(labelValues = ["discv5"])
+
+  else:
+    disc_incoming_packets.inc(labelValues = ["discv4"])
 
 func eligibleNode(proto: Eth1Discovery, rec: Record): bool =
   # Filter out non `eth` node
@@ -198,15 +210,18 @@ proc lookupRandomNode*(proto: Eth1Discovery, queue: AsyncQueue[NodeV4]) {.async:
   if proto.discv4.isNil.not:
     let nodes = await proto.discv4.lookupRandom()
     for node in nodes:
+      discv4_nodes_discovered.inc()
       await queue.addLast(node)
 
   if proto.discv5.isNil.not:
     let nodes = await proto.discv5.queryRandom()
     for node in nodes:
+      discv5_nodes_discovered.inc()
       if not proto.eligibleNode(node.record):
         continue
       let v4 = node.to(NodeV4).valueOr:
         continue
+      discv5_nodes_discovered.inc(labelValues = ["eligible"])
       await queue.addLast(v4)
 
 proc getRandomBootnode*(proto: Eth1Discovery): Opt[NodeV4] =

execution_chain/networking/p2p_metrics.nim

@@ -17,10 +17,10 @@ export
 
 declarePublicGauge rlpx_connected_peers, "Number of connected peers in the pool"
 
-declarePublicCounter rlpx_connect_success, "Number of successfull rlpx connects"
+declarePublicCounter rlpx_connect_success, "Number of successfull rlpx connects", labels = ["discversion"]
 
 declarePublicCounter rlpx_connect_failure,
-  "Number of rlpx connects that failed", labels = ["reason"]
+  "Number of rlpx connects that failed", labels = ["reason", "discversion"]
 
 declarePublicCounter rlpx_accept_success, "Number of successful rlpx accepted peers"
 

execution_chain/networking/peer_pool.nim

@@ -126,10 +126,17 @@ proc connect[Network](p: PeerPoolRef[Network], remote: Node): Future[PeerRef[Net
   # TODO: Probably should move all this logic to rlpx.nim
   if res.isOk():
     rlpx_connect_success.inc()
+    if remote.fromDiscv5:
+      rlpx_connect_success.inc(labelValues = ["discv5"])
+    else:
+      rlpx_connect_success.inc(labelValues = ["discv4"])
     return res.get()
   else:
     rlpx_connect_failure.inc()
-    rlpx_connect_failure.inc(labelValues = [$res.error])
+    if remote.fromDiscv5:
+      rlpx_connect_failure.inc(labelValues = [$res.error, "discv5"])
+    else:
+      rlpx_connect_failure.inc(labelValues = [$res.error, "discv4"])
     case res.error():
     of UselessRlpxPeerError:
       p.addSeen(remote.id, SeenTableTimeUselessPeer)