Network Latency

Overview

Meaning and scope

Network latency measures delay across peer links, RPC calls, validator messages, block propagation , relayers, or user connections. It includes transmission, routing, queueing, processing, and sometimes retry time. Latency varies by region, load, peer quality, payload size, and protocol design.

How it works

Median values can look healthy while tail latency causes timeouts or missed consensus duties. Low RPC latency also does not guarantee fast block inclusion or finality. Operators should monitor p50, p95, and p99 latency for each critical path and correlate spikes with errors and chain progress. Applications need realistic timeouts, cancellation semantics, and idempotent retries. Payment experiences should report pending status rather than interpreting a slow response as a failed or unpaid transaction.

Operational checks and limits

Network latency is the delay to transmit information between endpoints, not the total time for a payment to achieve finality. Measure its distribution and peaks rather than relying on one average.

Practical example

A remote node receives a new block three seconds later than its peers, delaying customer payment detection.

Distinctions

Latency measures transmission delay; block propagation is the spread of block information throughout the peer network.

How to evaluate it

A healthy payment service needs more than a single ‘network online’ indicator; for Network Latency, separate transaction propagation, peer connectivity, canonical-chain progress, block arrival, and the age of locally available data. When evaluating Network Latency, during an incident, retain observations from more than one independent endpoint; that helps distinguish a widespread protocol issue from a local network or monitoring failure without prematurely marking a customer payment as failed. Related concepts: Network; block propagation; block inclusion.