Progress: Scenario A-3 — operator dashboard, safety events, Config, and fleet polish in the private WebUI

[massivedev0]

Context

This is a progress update for scenario A (private self-hosted WebUI, directHttp transport) from the original proposal in Proposal: Modern graphical WebUI for ADAMANT TradeBot.

Scenario A is delivered in three increments. Increments #2 (fleet, market, commands, logs, 2FA) and #3 (live situation, MM health dots, params UX) are closed. This post covers increment 3 — the operator dashboard, safety stitching, admin Config, and fleet shell polish.

Tracking: adamant-tradebot-webui#1 (parent), adamant-tradebot-webui#4 (this increment, closed). Companion bot API work: adamant-tradebot#136.

Rule unchanged: the WebUI never connects to exchanges. All market, account, and parameter data flows from the bot /api/v1 API.

---

What shipped in scenario A-3

Operator dashboard

The per-bot Dashboard is now the primary landing view after you pick a bot tab:

• Overview — pair, exchange, MM state from GET /api/v1/status, 24h price/volume, and a feature edition badge (Basic → Custom → Advanced → Premium → Full) derived from installed bot capabilities, not from which modules happen to be ON right now
• Start / Stop MM — start no longer sends a strategy override; the bot keeps the current policy
• Compact price chart (15m default) with forming candle
• Modules grid — same order as Parameters, color-coded active/inactive
• Trading parameters summary — ON/OFF snapshot of every MM group
• Structured balances and open orders by type

Dashboard — one screen for “is MM running, what edition is this bot, and what is configured right now?”

Market polish

• Forming candle on the OHLC chart
• Bot-order markers on the order book and recent trades
• Open-order overlays on the chart
• Cancel sends { id, market, side }; already-gone orders are treated as success for the operator

Market — monitor liquidity placement, spread, and balances without leaving the WebUI

Parameters: live situation + safety

• 12-hour Live situation strip (inventory USD mix + open-order notional by purpose) in the operator timezone
• Hour frames from local health history: yellow = degraded hour, grey = MM stopped that hour
• Safety notices — when the bot auto-stops MM or pauses a ladder for safety, the WebUI surfaces a dialog from bot notify text; Parameters reflects emergency MM-off and ladder pause state

Parameters — edit strategy live; the situation strip shows how inventory and order mix evolved over the last 12 hours

Events board: audit + safety actions

• Parameter-change audit log (WebUI vs bot-side source) with balance snapshots at change time
• Safety badges (e.g. Auto-stopped for safety)
• Resume ladder action directly from a safety event row

Events — who changed what, when, with balance context; safety stops are visible and actionable

Config (admin)

• Top row: Runtime, Pair, and live Trade params (tradeParams_*.js tree) at equal height
• config.json Summary (Identity, Market, WebUI & API, Messaging, Modules, Limits) and raw JSON tabs
• Exchange report scrolls inside the Pair card; socket fields show Not included when the bot reports socketEnabled !== true

Config (admin) — runtime health, pair metadata, and a sanitized config snapshot without secrets

Fleet shell and registry

• Registry fields: editable Label (default {Exchange}-{tb|liq} from bot_id) and Account (default config.account); Transport, Version, Branch, Pair balance $
• Header tabs show {label} + {account}; offline rows are muted, connected-but-stopped bots stay full contrast
• Keep-alive panes — vertical scroll and in-page state survive bot tab switches; hidden panes read a frozen Redux snapshot instead of collapsing to EmptyState

---

What this means if you operate a token MM bot

Before (messenger-style WebUI)	After (scenario A, increment 3)
Text replies for status and params	Persistent dashboard with chart, modules, and params summary
Hard to see why MM stopped	Safety events with badges, balance context, and resume actions
Switching bots loses context	Multi-bot tabs with keep-alive panes and fleet overview
Config buried in files / CLI	Admin Config view with runtime health and sanitized config.json
“Is the bot actually working?” guesswork	MM-quality dots (working / degraded / stopped) from /status with a 10-minute grace period after start

You still self-host everything: WebUI behind your reverse proxy, bots on your infrastructure, no exchange API keys in the WebUI, mandatory 2FA for operators (email / ADM / ETH wallet).

---

Stack reminder

• WebUI v0.2.0 — Vite + React 18 + Chakra UI, Fastify BFF, SQLite event log
• Bot — GET /api/v1/bot, /status, /params, /account/*, WS params:updated
• Transport — DirectHttpTransport only (scenario B relay is not started)

Repo: Adamant-im/adamant-tradebot-webui

---

What is next

Scenario B (public subscription WebUI via outbound relay + license token) remains on the parent issue #1. Scenario A private self-host is feature-complete for the MVP scope defined in discussion #49; further A work is polish and bot API alignment, not new deployment models.

Questions and feedback welcome — especially from token teams running the free bot today and considering the private WebUI edition.