Firo Spark Sync Monitor

Where delivery stands

Live probes, Spark setHash health, durable uptime, and the public JSON API run on firostatus.com. M1 public gates (domain + HTTPS, public MIT GitHub, forum launch) are marked complete. Remaining focus is M2 ops (alerting, operator self-check, upstream, handover) so wallet teams can decide, not only watch a status wall.

M1 acceptance checklist · live

CFC verifies product signals here plus the three public gates below. Scorecard and compare are early M2 surfaces.

• …
  anonset_source = live

  Always-on sweeps (not Vercel snapshot)
• …
  Anon-set fields populated

  ms / MB / setHash on endpoints
• …
  /api/history recording

  sample_count + uptime %
• ok
  Own domain + HTTPS

  firostatus.com · always-on
• ok
  Public MIT GitHub

  README · methodology · CONTRIBUTING
• ok
  Forum launch write-up

  point at live URL + API curls

# Verifier cheatsheet (pass when live)
curl -sS ORIGIN/api/status | jq '{anonset_source, spark_ok:(.endpoints|map(select(.spark_consistent==false or (.anonset.consistent==false)))|length==0), sample_hint:(.endpoints|map(.anonset.ms)|map(select(.!=null))|length)}'
curl -sS ORIGIN/api/history?hours=24 | jq '{sample_count, fleet_pct_24h:.uptime.fleet_pct_24h, fleet_n:(.fleet|length)}'
curl -sS -f ORIGIN/api/ci | jq '{ok,spark_ok,max_lag,reasons}'

Who this is for

Primary: wallet / SDK / mobile teams reviewing default Electrum hosts, and Electrumx operators watching Spark serve cost. Secondary: public transparency when Spark sync stalls. The dashboard and the JSON feed show the same probe data.

Where this sits in the Firo stack

curl -sS 'ORIGIN/api/history?hours=24&limit=200' \
  | jq '{fleet_n:(.fleet|length), events:(.events.sethash|length),
         uptime:.uptime.fleet_pct_24h,
         last:(.points[-1]|{t,id,status,lag,anonset_ms,setHash,tls_days_left})}'

Public JSON routes

Fields that matter in /api/status

Height only checks vs Spark path

Concrete jobs this API supports

1. CI or ops alert. Fail a job or page someone when any default host has spark_consistent === false or anonset.consistent === false.
2. Incident triage. Separate an app bug from an Electrum host that is serving a slow or divergent Spark set.
3. Default list review. Compare anonset.mb and anonset.ms across operators before publishing recommended backends (same setHash can still mean different bytes on the wire).
4. Docs transparency. Embed /api/badge or poll /api/ci from a status page.

Example: list Spark inconsistent hosts

curl -sS ORIGIN/api/status \
  | jq '[.endpoints[]
      | select(.spark_consistent == false
          or (.anonset != null and .anonset.consistent == false))
      | {name, lag, spark_consistent, anonset}]'

Example: spark_ok-first CI (recommended for Spark sync)

# Prefer spark_ok for Spark path health (tip lag alone should not fail Spark CI)
curl -sS ORIGIN/api/ci | jq -e '.spark_ok == true'
# Strict tip+Spark gate (HTTP 503 when ok=false):
curl -sS -f ORIGIN/api/ci | jq '{ok, spark_ok, max_lag, reasons}'

ok requires spark_ok and max_lag ≤ 2. Mild lag alone can yield spark_ok: true with ok: false. Use spark_ok for Spark-only gates; use -f + ok when tip freshness is also required.

Example: compact consensus poll

curl -sS ORIGIN/api/spark \
  | jq '{checked_at, spark_consensus, spark_sethash_consensus, summary, inconsistent}'

Pin / audit fleet setHash

# Consensus fingerprint
curl -sS ORIGIN/api/status | jq -r '.spark_sethash_consensus'
# Hosts that disagree
curl -sS ORIGIN/api/status | jq '[.endpoints[] | select(.anonset.consistent==false) | {id,name,setHash:.anonset.setHash}]'
# Divergence windows (always-on)
curl -sS 'ORIGIN/api/history?hours=168' | jq '.events.sethash[-5:]'

Embed kit

<!-- README / status page -->
![Spark fleet](ORIGIN/api/badge)

<!-- Deep links -->
Overview: ORIGIN/overview
Spark health: ORIGIN/spark
Backend detail: ORIGIN/backend/mathnodes
Developers: ORIGIN/developers

Registry ids and deep links also ship in /api/docs (registry, deep_links).

ElectrumX method → JSON field

Privacy and probe surface

Probes are public Electrum methods only: server.version, blockchain.headers.subscribe, spark.getsparklatestcoinid, and spark.getsparkanonymityset. No addresses, keys, transactions, or wallet traffic are sent or stored. CORS is open (*) so browser tools and dashboards can read the feed.

Start here: /api/docs · /api/status · /api/ci

What this is

A community health board and JSON feed for the public Firo Electrum backends that light and mobile wallets depend on. Flagship focus: Spark sync health (anon set fetch and setHash), plus tip lag. The dashboard and the API show the same data. See Developers for integration details.

How to read the Overview in 30 seconds

1. Live strip — gates, risk, agreement, then durable charts (lag / fleet mix / anon-set).
2. Fleet at a glance + server history — who is lagging or down over 24h/7d.
3. Scorecard → status changes → Spark signal / setHash incidents; framing question at the bottom.

What it checks (privacy safe)

• server.version: reachability, TLS, software version, cert expiry
• blockchain.headers.subscribe: chain height and lag vs fleet reference tip
• spark.getsparklatestcoinid: active Spark coin group id, cross checked across backends
• spark.getsparkanonymityset: full anon set fetch time, size, and setHash consistency

No addresses, keys, transactions, or wallet data are sent or stored. The reference tip is the highest height at least two backends agree on. Spark consensus is the modal value across the fleet.

Status model

• green: reachable, ≤2 blocks behind, Spark id and setHash match fleet, probe RTT under 3s, TLS cert ≥14 days
• yellow: lagging, slow RTT, missing Spark id, Spark coin id / setHash mismatch, or TLS <14 days — tip can be 0 blk with Spark id matching and still yellow on setHash ✗
• red: unreachable, more than 100 blocks behind, or TLS certificate expired

Always-on vs preview

Privacy statement

Probes are public Electrum methods only. No addresses, keys, transactions, wallet traffic, or personal data are collected. This board is ops health for the curated fleet — not a wallet and not a privacy guarantee for funds.

Scoring (short)

Reference tip = highest height ≥2 backends agree on. Uptime = strict green % (yellow/red count as down). Anon-set ms/MB are from this monitor’s single region — not global wallet RTT. Methodology lives in docs/METHODOLOGY.md once the repo is public.

Public JSON API: /api/status · /api/history · field docs: /api/docs · badge: /api/badge