Details
Ops Dashboard — pipeline, rollout ladder, and incident board
Operating-loop health, where each bot sits on the rollout ladder, and live incidents. Switch to Trader Home for PnL, edge, and blotter.
Discover
28 markets · 3 new today
Ingest
9 feeds · 0 gaps/min
State
28 books · 0 stale
Signals
142 features/s
Fair value
2 suppressed (low conf)
Execution
q_exec err 3.1 bps
Risk
1 cluster 96% of cap
Place
1,284 env · 4.2% reject
Reconcile
0 unmatched
Replay
81
historical only
Shadow
19
live read, no orders
Paper
0
simulated fills
Canary
9
demo-wired
Broader live
0
production-live
Bots advance one rung at a time. Every promotion requires 27/27 sections + green replay parity. 0 are in production-live — demo-wired ≠ production-live.
-
auto-paused
Feed gap on Polymarket WS ·
BBO_UPDATEDdropped 2s -
resync queued
Book stale on
ai-frontier-model-2026· 22s without update -
rollback pending
Calibration drift on
fair-value-v1.4· Brier +0.08 vs baseline -
operator review
New market discovered ·
uk-snap-election-2026
Stale books
0 / 28
threshold 15s
Fill ratio (1h)
94.2%
target ≥ 92%
Slippage err
3.1 bps
predicted vs actual
Heartbeats
9 / 9
all wired bots live
Envelopes
1,284
today · peak 18.4/min
Reject rate
4.2%
down from 5.1%
RPC latency
182 ms
p95 last 5m
Hash parity
100%
live vs replay
Why this view exists
Ops Dashboard answers “is the machine healthy?”. The operating-loop strip up top maps to the 9-stage pipeline from the explainer (PDF pp 1–2). The rollout ladder (PDF pg 14) shows how many bots are at each rung — promotion requires 27/27 sections + green replay parity. The incident board surfaces auto-paused feeds, stale books, and calibration drift with their reason codes. Monitoring tiles cover the plumbing: stale books, fill ratio, slippage error, heartbeats, envelope throughput, reject rate. All numbers synthetic — not a forecast.