Data · statusRefreshes every 30 s · UTC

Data quality is public system state.

Per venue and per collector, with the age of every figure. Overall state is shown as the service states it, beside the counts it summarises.

There is no single "all systems operational" badge. A failing collector on one venue is shown as exactly that, next to the venues that are collecting normally.

Overall · GET /v1/health Time of the last response, printed from the response

Requesting

Only venues listed by the coverage response are shown; internal service segments in the health response are hidden.

Collectors

One row per venue. Open it for one row per collector.

A venue collects through several collectors — snapshot, depth, funding, candles, trades, liquidations — and they fail independently. The venue row carries the worst state among them; the rows inside carry their own.

Collectors · per venue · GET /v1/health Expandable rows · ages beside every timestamp

Venue	Collector state	Last success · age	Consecutive failures	Last error · age	Instruments stale
Requesting

Error text is shortened. No stack trace and no internal address is shown. Status is never colour alone — every chip carries its word.

Stale instruments

Instruments whose latest snapshot is old.

An instrument no longer collected still has a stored row. It is listed here with the age of its latest snapshot, and it appears in Arena as STALE — never as a live price.

Stale · from the same responseGrouped by venue · symbols as the venue spells them

Requesting

Coverage

The record behind the present state.

What each venue collects, since when, and the hours nothing was observed.

See coverage

Methodology

How age and state are computed.

Thresholds per call group, and what LIVE, DELAYED and STALE mean.

Read the methodology