initial jaeger integration (#2047)

Co-authored-by: Pierre Krieger <pierre.krieger1708@gmail.com>
Co-authored-by: Peter Goodspeed-Niklaus <coriolinus@users.noreply.github.com>
Co-authored-by: Bastian Köcher <bkchr@users.noreply.github.com>

polkadot/node/network/availability-distribution/src/lib.rs

@@ -44,6 +44,7 @@ use polkadot_subsystem::messages::{
 	NetworkBridgeMessage, RuntimeApiMessage, RuntimeApiRequest,
 };
 use polkadot_subsystem::{
+	jaeger,
 	errors::{ChainApiError, RuntimeApiError},
 	ActiveLeavesUpdate, FromOverseer, OverseerSignal, SpawnedSubsystem, Subsystem,
 	SubsystemContext, SubsystemError,
@@ -354,6 +355,8 @@ where
 				}
 			};
 
+			let mut _span = jaeger::hash_span(&gossiped_availability.candidate_hash.0, "availability-message-received");
+
 			process_incoming_peer_message(ctx, state, remote, gossiped_availability, metrics)
 				.await?;
 		}

polkadot/node/network/bitfield-distribution/src/lib.rs

@@ -27,6 +27,7 @@ use futures::{channel::oneshot, FutureExt};
 
 use polkadot_subsystem::messages::*;
 use polkadot_subsystem::{
+	jaeger,
 	ActiveLeavesUpdate, FromOverseer, OverseerSignal, SpawnedSubsystem, Subsystem, SubsystemContext, SubsystemResult,
 };
 use polkadot_node_subsystem_util::metrics::{self, prometheus};
@@ -180,7 +181,9 @@ impl BitfieldDistribution {
 
 					for relay_parent in activated {
 						tracing::trace!(target: LOG_TARGET, relay_parent = %relay_parent, "activated");
-						// query basic system parameters once
+						let _span = jaeger::hash_span(&relay_parent, "bitfield-dist:active_leaves:basics");
+
+						// query validator set and signing context per relay_parent once only
 						match query_basics(&mut ctx, relay_parent).await {
 							Ok(Some((validator_set, signing_context))) => {
 								// If our runtime API fails, we don't take down the node,
@@ -232,6 +235,7 @@ where
 	Context: SubsystemContext<Message = BitfieldDistributionMessage>,
 {
 	tracing::trace!(target: LOG_TARGET, rep = ?rep, peer_id = %peer, "reputation change");
+
 	ctx.send_message(AllMessages::NetworkBridge(
 		NetworkBridgeMessage::ReportPeer(peer, rep),
 	))
@@ -306,6 +310,9 @@ async fn relay_message<Context>(
 where
 	Context: SubsystemContext<Message = BitfieldDistributionMessage>,
 {
+	let span = jaeger::hash_span(&message.relay_parent, "relay_msg");
+
+	let _span = span.child("provisionable");
 	// notify the overseer about a new and valid signed bitfield
 	ctx.send_message(AllMessages::Provisioner(
 		ProvisionerMessage::ProvisionableData(
@@ -318,6 +325,9 @@ where
 	))
 	.await;
 
+	drop(_span);
+
+	let _span = span.child("interested peers");
 	// pass on the bitfield distribution to all interested peers
 	let interested_peers = peer_views
 		.iter()
@@ -341,6 +351,7 @@ where
 			}
 		})
 		.collect::<Vec<PeerId>>();
+	drop(_span);
 
 	if interested_peers.is_empty() {
 		tracing::trace!(
@@ -349,6 +360,7 @@ where
 			"no peers are interested in gossip for relay parent",
 		);
 	} else {
+		let _span = span.child("gossip");
 		ctx.send_message(AllMessages::NetworkBridge(
 			NetworkBridgeMessage::SendValidationMessage(
 				interested_peers,
@@ -483,6 +495,8 @@ where
 		NetworkBridgeEvent::PeerMessage(remote, message) => {
 			match message {
 				protocol_v1::BitfieldDistributionMessage::Bitfield(relay_parent, bitfield) => {
+					let mut _span = jaeger::hash_span(&relay_parent, "bitfield-gossip-received");
+					_span.add_string_tag("peer-id", &remote.to_base58());
 					tracing::trace!(target: LOG_TARGET, peer_id = %remote, "received bitfield gossip from peer");
 					let gossiped_bitfield = BitfieldGossipMessage {
 						relay_parent,
@@ -581,6 +595,8 @@ where
 		return;
 	};
 
+	let _span = jaeger::hash_span(&message.relay_parent, "gossip");
+
 	job_data.message_sent_to_peer
 		.entry(dest.clone())
 		.or_default()

polkadot/node/network/collator-protocol/src/collator_side.rs

@@ -24,6 +24,7 @@ use polkadot_primitives::v1::{
 	CollatorId, CoreIndex, CoreState, Hash, Id as ParaId, CandidateReceipt, PoV, ValidatorId,
 };
 use polkadot_subsystem::{
+	jaeger,
 	FromOverseer, OverseerSignal, SubsystemContext,
 	messages::{
 		AllMessages, CollatorProtocolMessage,
@@ -430,6 +431,8 @@ async fn process_msg(
 			state.collating_on = Some(id);
 		}
 		DistributeCollation(receipt, pov) => {
+			let _span1 = jaeger::hash_span(&receipt.descriptor.relay_parent, "distributing-collation");
+			let _span2 = jaeger::pov_span(&pov, "distributing-collation");
 			match state.collating_on {
 				Some(id) if receipt.descriptor.para_id != id => {
 					// If the ParaId of a collation requested to be distributed does not match
@@ -539,10 +542,12 @@ async fn handle_incoming_peer_message(
 			);
 		}
 		RequestCollation(request_id, relay_parent, para_id) => {
+			let _span = jaeger::hash_span(&relay_parent, "rx-collation-request");
 			match state.collating_on {
 				Some(our_para_id) => {
 					if our_para_id == para_id {
 						if let Some(collation) = state.collations.get(&relay_parent).cloned() {
+							let _span = _span.child("sending");
 							send_collation(ctx, state, request_id, origin, collation.0, collation.1).await;
 						}
 					} else {

polkadot/node/network/collator-protocol/src/validator_side.rs

@@ -30,6 +30,7 @@ use polkadot_primitives::v1::{
 	Id as ParaId, CandidateReceipt, CollatorId, Hash, PoV,
 };
 use polkadot_subsystem::{
+	jaeger,
 	FromOverseer, OverseerSignal, SubsystemContext,
 	messages::{
 		AllMessages, CandidateSelectionMessage, CollatorProtocolMessage, NetworkBridgeMessage,
@@ -504,6 +505,7 @@ where
 			state.peer_views.entry(origin).or_default();
 		}
 		AdvertiseCollation(relay_parent, para_id) => {
+			let _span = jaeger::hash_span(&relay_parent, "advertising-collation");
 			state.advertisements.entry(origin.clone()).or_default().insert((para_id, relay_parent));
 
 			if let Some(collator) = state.known_collators.get(&origin) {
@@ -515,6 +517,8 @@ where
 			modify_reputation(ctx, origin, COST_UNEXPECTED_MESSAGE).await;
 		}
 		Collation(request_id, receipt, pov) => {
+			let _span1 = jaeger::hash_span(&receipt.descriptor.relay_parent, "received-collation");
+			let _span2 = jaeger::pov_span(&pov, "received-collation");
 			received_collation(ctx, state, origin, request_id, receipt, pov).await;
 		}
 	}
@@ -659,6 +663,7 @@ where
 			);
 		}
 		FetchCollation(relay_parent, collator_id, para_id, tx) => {
+			let _span = jaeger::hash_span(&relay_parent, "fetching-collation");
 			fetch_collation(ctx, state, relay_parent, collator_id, para_id, tx).await;
 		}
 		ReportCollator(id) => {

polkadot/node/network/statement-distribution/src/lib.rs

@@ -23,6 +23,7 @@
 #![warn(missing_docs)]
 
 use polkadot_subsystem::{
+	jaeger,
 	Subsystem, SubsystemResult, SubsystemContext, SpawnedSubsystem,
 	ActiveLeavesUpdate, FromOverseer, OverseerSignal,
 	messages::{
@@ -828,6 +829,7 @@ async fn handle_network_update(
 					).await;
 
 					if let Some((relay_parent, new)) = new_stored {
+						let mut _span = jaeger::hash_span(&relay_parent, "sending-statement");
 						// When we receive a new message from a peer, we forward it to the
 						// candidate backing subsystem.
 						let message = AllMessages::CandidateBacking(
@@ -943,6 +945,7 @@ impl StatementDistribution {
 				FromOverseer::Communication { msg } => match msg {
 					StatementDistributionMessage::Share(relay_parent, statement) => {
 						let _timer = metrics.time_share();
+						let mut _span = jaeger::hash_span(&relay_parent, "circulate-statement");
 
 						inform_statement_listeners(
 							&statement,