Downgrade routine peer interaction DEBUG logs to TRACE in Discovery v5 handlers

[parthdagia05]

Summary

Discovery v5 handler classes log routine peer-interaction failures (bad packets, failed handshakes, rejected peers) at DEBUG. On a live Ethereum network these messages fire hundreds of times per minute, making --logging=DEBUG output difficult to read when investigating unrelated issues.

This issue proposes downgrading 16 LOG.debug() calls across 9 files in org.ethereum.beacon.discovery.pipeline.* and ...schema.NodeSession to LOG.trace(). The messages remain available for deep protocol debugging they are simply no longer part of the default DEBUG stream.

Downstream discussion: besu-eth/besu#9691 "Tune Discovery v5 Logging Levels".

Measured Impact

We ran Besu 25.3.0 (pre-built release) against Ethereum mainnet for 5 minutes with --logging=DEBUG and captured the following:

Level	Lines	Per minute
Total	18,023	~3,605
DEBUG	8,029	~1,606 (~27/sec)
INFO	67	~13
WARN	3	<1
ERROR	1	<1

Top DEBUG sources

Class	Count	Notes
RlpxAgent	2,054	(out of scope for this repo)
EthPeers	1,254	(out of scope)
DeFramer	1,022	(out of scope)
RecursivePeerRefreshState	548	Discovery-related
PeerDiscoveryController	479	Discovery-related
VertxPeerDiscoveryAgent	85	Discovery-related

Discovery-specific noise (originating in this repo's handlers)

Message	Count	Per minute
"Handshake timed out"	469	~94
"Peer is unreachable"	60	~12
"Discarding invalid packet"	25	~5
Total discovery DEBUG	~570	~114

Note: This capture was taken on a home machine with limited peer connectivity. A production mainnet node maintaining 50+ active peers will produce substantially higher volumes these numbers are a lower bound.

Affected Files

All 16 statements describe peer-caused events that our node handled correctly bad bytes, invalid ENRs, policy-rejected addresses, protocol violations, or sessions that timed out because a peer went silent. None indicate a local bug.

File	Count	Nature of messages
pipeline/handler/HandshakeMessagePacketHandler.java	6	Missing WhoAreYou challenge, invalid ENR, wrong node ID, disallowed ENR, bad ID signature, failed message read
pipeline/handler/WhoAreYouPacketHandler.java	2	Nonce verification failed, failed to read message
pipeline/info/FindNodeResponseHandler.java	2	Rejecting invalid record, wrong distance in response
schema/NodeSession.java	1	Cancelling requests after peer timeout
pipeline/handler/BadPacketHandler.java	1	Bad packet received
pipeline/handler/OutgoingParcelHandler.java	1	Dropping packet to disallowed destination
pipeline/handler/UnauthorizedMessagePacketHandler.java	1	Failed to read unauthorized message
pipeline/handler/MessagePacketHandler.java	1	Failed to read message
pipeline/handler/PacketSourceFilter.java	1	Ignoring disallowed source
Total	16

Proposed Fix

Change each of the 16 LOG.debug(...) calls above to LOG.trace(...). This:

• Removes ~114+ messages per minute from the default DEBUG stream, making Besu's DEBUG logs usable again for troubleshooting unrelated issues.
• Preserves every message for developers doing protocol-level debugging --logging=TRACE still shows everything.
• Does not modify any LOG.warn(), LOG.error(), or LOG.info() calls.
• All existing tests continue to pass.

Deliberately Out of Scope

Two additional LOG.debug() calls in schema/NodeSession.java were considered but intentionally left at DEBUG:

• NodeSession#cancelAllRequests logging an exception thrown while completing a cancelled request's promise.
• NodeSession#cancelAllRequests logging when clearRequestInfo returns null for a requestId we just iterated from the same map.

Both signal potential local code issues (an unexpected exception in our own cleanup path, or a possible race on requestIdStatuses) rather than routine peer behavior, so they stay visible at DEBUG for operators troubleshooting their own node.

Status

A PR with the 16-statement change is ready and will be opened once this issue is filed for discussion.