Status and stats
Read service checks, diagnose a waiting flow and understand what the public activity totals count.
Open Status from More → About. Its headline is All systems live, Running degraded or Venue unreachable. A degraded headline does not identify the failed feature by itself; open the affected system's group and read its checks, details and last-checked time.
Inspect the actual degraded Status page, its Projector and Ops checks, then navigate directly to Stats. Public-call and qualified-Window counters have a defined scope. The independently inspected recount was older than the live session.
0:28 · Captured 8 October 2026 · Silent video with captions.
Read the transcript
0:00 · Read the degraded status. Read the real Running degraded status. This capture shows eight failing checks, two slow, twenty-three healthy and three marked off-hours or not set up; it does not establish a healthy venue.
0:06 · Inspect the Projector. Open the Projector disclosure and inspect that component’s reported check. A healthy component does not make the entire venue healthy.
0:12 · Inspect Ops. Open the Ops disclosure and inspect its individual checks.
0:15 · Open public Stats. Navigate directly to the public Stats URL and wait for the actual page load.
0:18 · Read proof of demand. The public Proof of demand counter shows zero published calls. That public count is not a count of all private trades.
0:20 · Read venue totals. Read the venue totals with their qualified-Window scope: five or more traders. Zero qualified totals do not mean no private trading activity.
0:24 · Find the auditor view. The Auditor view begins below venue totals; its details extend below this clip’s viewport. The separately inspected recount was timestamped 16:41:43 UTC at offset 2,458,892, older than this session near 18:00 UTC, so it does not verify every current contract.
Find the relevant check
| What you see | Where to look |
|---|---|
| No party or credits after taking a seat | Guest-seat pool and seat-related ops checks |
| No Window or quote | Roller, pricer, venue mode, paused lanes and the asset's source |
| Closed Window without a result | Oracle freshness and resolver backlog |
| Result without a payout | Settler backlog and projector lag |
| Old balance or history | Ledger end, projection cursor and database |
| Assistant unavailable | Sensei's model check |
Hosted check — 8 October 2026, 18:00 UTC: the final recording showed Running degraded, with failing oracle freshness and a resolver backlog. Print-source checks reported missing-print voids, RedStone was late, and Pyth's OpenAI and Anthropic valuation requests reported 403 / not entitled. Ops, projector, settler and guest-seat checks displayed healthy. Those individual checks did not establish overall health; the resolver had been healthy in an earlier read and later fell behind. Expand the current group for its latest result.
The page refreshes every 30 seconds. It distinguishes a configured service failing to answer from an optional service not set up here. A stock source closed for its session can be expected; it is not automatically an outage. A healthy read is a point-in-time check, not proof that your submission succeeded. Check the position or transaction receipt too.
Read Stats
Open Stats from More → Records and proof. Adoption and live activity count calls their owners chose to publish. Private trading is not included. Activity rows link to ledger updates.
Venue totals use a privacy threshold: Windows with fewer than five traders have their trading figures withheld. A paging-limited read labels its figures as floors. Those totals are not a complete count of every private trader or call. In the final 8 October hosted check, Stats showed zero published calls and zero trades in qualifying Windows with at least five traders, despite confirmed seat fills. Those filtered zeroes did not mean no trading occurred.
The auditor section separates a reserve snapshot from an independent recount. Read its timestamp and result; no recount yet is different from agreement. In the 18:00 UTC recording, the displayed independent recount was from 16:41:43 UTC. It was a comparison at that older offset, not a fresh attestation of every current value. These are test-network credit figures, not external cash or evidence of public adoption beyond the calls counted.