lahza.xyz

Time is the only
honest metric.

Lahza listens to every Avalanche L1 from inside the network — three observatories on three continents timing every user transaction with their own clocks — and publishes a score no one can buy.

every mainnet L13 observatories · every transaction0 transactions sent — we only listen
scroll

Every chain grades its own homework.

Each L1 monitors itself, from the inside. A node can report "healthy" while users wait ten seconds for a transaction. Institutions moving billions onto these chains ask one question first: who independently verifies this? Today, nobody does.

frankfurt → meridian-finance
0.000
true finality, measured from outside

You are riding a real user transaction we just observed. This is exactly how we measure every chain: see it enter the network, follow it through consensus, stop the clock the instant it finalizes.

01Listen to every chain

Our own nodes join every mainnet L1 we cover — blocks, validator sets, ICM traffic, and every transaction as it arrives. Passive, continuous, free for every chain.

watching 100% of mainnet L1s

02Measure every transaction

Observer nodes in Frankfurt, Virginia and Singapore sit inside each network and stamp every user transaction three times — first seen, included, final — with their own clocks. Nothing is sent; everything is measured. None of them run on infrastructure of any chain we observe.

tx seen 14:02:09.113 · final +441ms

03Tell teams and the world

Teams get an alert within seconds of degradation. The world gets a public score every chain can be judged — and trusted — by.

push → websocket, webhook, api · public → x
Public scorecard

One number every chain answers to.

Measured finality, uptime, validator health and ICM delivery — blended into a single grade, published openly, computed the same way for everyone.

live network view
ChainScoreFinality p50Finality p99Uptime 90dICM delivery
meridian-finance99.40.41s0.92s100%100%
aurora-gaming98.70.47s1.18s99.98%99.7%
polar-assets97.90.52s1.31s99.95%99.2%
helios-play94.10.88s2.60s99.71%97.4%
northline88.31.42s4.10s99.12%93.8%
fictional demo chains · in production every figure traces back to a real user transaction anyone can verify
best finalitymeridian-finance most stablepolar-assets most consistent across regionsaurora-gaming best message deliverymeridian-finance
0.61s
Avalanche Reliability Index Median measured finality across every mainnet L1, published monthly — the one number the ecosystem can quote. Demo value.
The blind spot

Cross-chain messages fail silently. We watch every one.

Value moves between L1s through Interchain Messaging. When a message stalls, no dashboard anywhere shows it. Lahza times every cross-chain message it observes end to end — how long delivery takes and how often it lands — and raises the alarm on the first one that doesn't arrive. No other tool in the ecosystem watches this today.

Three-leg timing — source finality, relayer pickup, destination delivery
Stalled-message alert the moment it happens
Delivery-rate history per chain pair — and which leg breaks most
The instrument

More than a score.

Every number on the board opens into the evidence behind it. The public sees the picture; operators get the stream.

01Historical recordFinality and uptime over 24 hours, 7 days, 30 days. "It got worse last week" becomes a fact.
02Incident timelineWhen it broke, how long, when it recovered. Silent outages, made visible.
03Region gapThe same chain timed from three places — and the gap between them, scored.
04Three-leg message routesSource finality, relayer pickup, destination delivery — each leg timed, so a stall has an address.
05Live stream & pushWebSocket, signed webhooks, REST API. We push; you route.
06Signed scorecardA per-chain report an institution can cite: p50/p95, uptime, consistency, delivery, trend.
The score is public and is not sold. Depth — streams, raw measurements, breakdowns — comes through the ecosystem's partnership.
Research

Measurement you can cite.

"Sub-second finality" should be a measured fact, not a marketing line. Lahza publishes the first open, longitudinal dataset of real-world finality, validator churn and cross-chain latency across sovereign Avalanche L1s — written up for peer review.

Evidence on-chainEvery incident is published with the real user transaction that proves it; the raw record is committed on-chain every hour. Raw data is available to operators, licensees and auditors.
3-region methodFrankfurt, Virginia and Singapore observatories remove single-vantage bias from every figure — and none run on infrastructure of an observed chain.
Built for peer reviewMethodology and dataset written up as an academic measurement study of Avalanche consensus in the wild.
The model

The score is not for sale.

Two layers. Public: scores, incidents with their evidence transactions, the index — as a page, a feed and an API, for everyone. Evidence: raw measurements, three-leg timings, history, webhooks — opened to an ecosystem through its foundation's partnership. Institutions license the data.

3observatories 0transactions sent 3timestamps per transaction 1public formula

Public

Free
for everyone, always
  • Public dashboard — every mainnet L1, scored
  • Avalanche Reliability Index, published monthly
  • Public stream, API and incident feed on X
Open the dashboard

Foundation partnership

Annual agreement
for the foundation behind an ecosystem
  • Coverage: every mainnet L1, from three observatories
  • Monthly index and quarterly signed ecosystem report
  • The evidence layer opened to the whole ecosystem, in the foundation's name
  • A pool of operator seats for the ecosystem's L1 teams
For foundations

Operators

By request
for teams running an L1
  • Evidence access comes with your ecosystem's partnership — raw streams, webhooks, breakdowns
  • Incident detail with three-leg ICM breakdown, region breakdown, 30-day history
  • Not partnered yet? Request early access. No price list — chains never pay to be measured
Request early access

No agreement buys influence over a number. The public score stays free and identical for every chain, and no chain pays to be measured.

Give your ecosystem an independent stopwatch.

One agreement. Every chain measured. Published for everyone.

Get in touch