ProbeLab's Sonar tests whether an Ethereum node is reachable from outside
ProbeLab introduced Sonar on Ethereum Research on 19 November, a free tool, now in beta, that probes one Ethereum node on request. The user supplies an ENR, a libp2p peer ID or an enode URL. Sonar looks the node up over discv5, which also refreshes an outdated record, then dials every address the record advertises, IPv4 and IPv6, TCP and QUIC, one at a time and reports each separately with connection time, round-trip time and the full error when a dial fails.
The tool then identifies the node: consensus-layer clients through libp2p identify, execution-layer clients through the devp2p hello over RLPx. For consensus nodes it also requests Status and MetaData, records the gossipsub topics the node subscribes to and any GOODBYE message with its reason code, such as 129 for too many peers. It reads the node's discv5 routing table bucket by bucket and checks whether the client's API and metrics ports are open. A probe takes from a few seconds to about half a minute.
The post lists the problems this is meant to catch, which a node cannot see in its own logs: an ENR advertising an IP or port that is not reachable because of NAT, a firewall or a changed address; only some transports working, for example TCP but not QUIC; a mismatch between the ENR and the node's Status or MetaData in fork digest, attestation or sync-committee subnets or custody group count; and a beacon API or metrics port exposed to the internet. Results are compared with the node's own record and with ProbeLab's crawler history. ProbeLab is asking client teams and operators what else a probe should check.

Why it matters
A node that looks healthy locally can still be hard for peers to reach, and the operator has no view from the other side. An outside probe that a single operator can run on demand turns that blind spot into a checklist.