MarginPad Status - API, trading store, liquidation feed uptime
MARGINPAD

System status

What a bot depends on, checked every 10 minutes: the paper-trading store, the liquidation collector and the price floor. Days are UTC. Nothing here needs a login.

Checking…

Trading store

90 days agotoday

Liquidation collector

90 days agotoday

Price floor

90 days agotoday

Liquidation sources live from the collector

VenueSocketLast liquidation
Loading…

Each day shows the share of checks that passed. A day drawn pale is one where fewer checks ran than the every-10-minute schedule expects - the scheduler itself missed passes - so there is less evidence about that day; hover it for how many ran. We show it rather than hiding it, but a day whose every check passed is not counted against the service because we looked less often. The price floor is not the price endpoints. It is a cached copy of every price, rewritten once a minute, that a cold Cloudflare location serves instantly instead of waiting on the exchanges; a failed check means that rewrite stopped happening, so those requests paid the live round trip - slower, not unavailable. GET /api/v1/price answers either way, and its own latency is on the API page. Thin venues (dYdX, Bitfinex) legitimately go hours without a liquidation; a connected socket with an old last event is quiet, not broken. Machine-readable: GET /api/status, external monitors can use GET /api/health (200 while the scheduler is alive). Changes to the API: RSS.