The dashboard

The dashboard is Forven's home page — the route at /, the first screen after you sign in. It is the single pane where the whole lab reports for duty: system health, the day's task counts, any critical alerts, your equity, and the strategies currently in flight. It is built for a glance, not a study session. If something needs your attention, the dashboard is where it surfaces first.

This page explains every panel and, more importantly, how to read the status indicators so a green light and a red light each mean something specific.

What the dashboard is for

You open the dashboard to answer one question: is the lab healthy, and does anything need me right now? Everything on it serves that question. The KPIs tell you whether work is flowing. The alert banner tells you whether work has broken. The equity and trading panels tell you what is happening with positions. The leaderboard tells you which strategies are carrying the lab.

It is a read-and-route screen. You confirm state here, then click through to the page that owns the detail — risk, the strategy lab, live trading, or the operations console.

The panels

The dashboard composes several panels. Exactly which appear depends on your tier and what is running, but the core set is consistent.

System health and KPIs

A header strip reports the lab's vital signs and the day's throughput KPIs:

• Tasks completed today and tasks errored today — the work the agents and scheduler have processed since midnight. A rising error count is your earliest signal that something upstream is stuck.
• Active agents — how many agent roles are currently working.
• Jobs due next hour — scheduled work about to run, drawn from the scheduler. Useful for knowing whether a quiet lab is idle or simply between cycles.

Health rolls up to a single status the same way the health monitor does elsewhere — a green / amber / red signal you can read without parsing the components behind it. Amber means degraded but running; red means a component needs attention.

Critical alerts banner

When something crosses a critical threshold, a banner pins to the top of the dashboard and stays there until the condition clears or you dismiss it. This is the one panel designed to interrupt you. Kill-switch activation, a stalled agent, and other critical events route here. Treat a red banner as a stop-and-look, not a notification you scroll past.

Equity overlay

An equity panel charts account value over time. In a beta build this reflects the paper book — simulated capital, no real orders — because beta hard-locks execution to paper. The number is an honest record of the simulation, not a forecast and not a live balance. See execution modes for why paper equity does not reconcile against any exchange.

Strategy leaderboard

The leaderboard ranks the strategies the lab is running so you can see, at a glance, which are leading and which are lagging. It is a summary view: click any row through to the lab strategy detail page for the full backtest history, gauntlet verdicts, and lifecycle timeline. The leaderboard tells you who to look at; the detail page tells you why.

Pipeline flow and agent heartbeat

A compact pipeline panel shows where strategies sit along the lifecycle — researching → backtesting → quick_screen → gauntlet → paper → live — so you can see the shape of the funnel without leaving home. (On the marketing site these stages are named screen / gauntlet / candidate / live, where "candidate" is the paper stage; the app uses the real stage names. See the pipeline for the full map.)

An agent heartbeat indicator shows whether the agent layer is alive and processing. A missed heartbeat surfaces here before it shows up as a backlog.

Trading and paper summaries

A live trading panel and a paper trading summary give you a thumbnail of current positions and session activity. In beta this is paper only. For open and closed positions, force-close, and the signal chart workspace, click through to live trading; for paper sessions and overlays, see paper trading.

The activity stream

Below the panels sits a collapsible activity stream — the full firehose of what the lab is doing, moment to moment. Leave it collapsed when you only need the headline state; expand it when you are diagnosing why a number moved. It is a live tail, not a permanent log; the durable audit lives in the task queue.

Reading the real-time indicators

The dashboard updates itself. You should rarely need to refresh the browser, and understanding how it refreshes tells you whether a stale-looking screen is a problem or just a quiet moment.

• Event-driven refresh. The dashboard subscribes to a real-time event feed and re-renders the relevant panel when an event fires. Promotions (strategy_promoted), kill-switch activation (kill_switch), and agent stalls (agent_stalled) push an update the instant they happen — you do not wait for a poll.
• Polling fallback. If the real-time connection drops, the dashboard falls back to polling and keeps updating until the connection returns. A brief gap is expected on reconnect, not a fault.
• Per-panel timeout. Each panel has its own load budget (an illustrative ~8-second timeout); one slow panel shows its own loading or error state without freezing the rest of the dashboard. A single stuck tile does not take the page down.

Colour, briefly

Status colour is consistent across Forven: green is healthy / normal, amber/yellow is degraded or a soft limit approaching, red is a hard condition that needs action — a tripped circuit breaker, an active kill-switch, a failed health check. When you see red on the dashboard, the page that owns that condition (usually risk or health) has the detail and the action button.

Steps: a daily dashboard sweep

A fast, repeatable way to read the lab each session:

1. Open the home page (/). Let the panels settle past their initial load.
2. Read the health status first. Green, move on; amber or red, follow it to health or ops before doing anything else.
3. Check the critical alerts banner. If a kill-switch or daily-loss halt is showing, go straight to risk — that takes priority over everything below.
4. Scan the KPIs: compare tasks errored today against tasks completed. A spiking error count means open the task queue and inspect the failures.
5. Glance at equity and the trading summary for unexpected movement. In beta this is paper; treat it as simulation evidence, not a balance.
6. Skim the leaderboard for any strategy that has moved sharply. Click through to its lab detail page if you want the why.
7. If nothing is red and the error count is flat, you are done — the lab is running and you have confirmed it in under a minute.

What you'll see

On a healthy lab: a green health signal, no alert banner (or only dismissible informational notices), a completed-tasks count comfortably ahead of errors, an equity line that matches your paper sessions, and a leaderboard of strategies sitting in their expected stages. When the real-time feed is live, counts and the activity stream tick forward on their own. When something is wrong, the dashboard does not stay quiet — the alert banner appears, a panel turns amber or red, and the affected KPI moves.

Caveats

• Beta is paper-locked. All equity, trading, and PnL shown on the dashboard reflect simulated paper execution. There is no live capital in a beta build.
• The dashboard summarizes; other pages own the truth. Every panel is a click-through to a page that holds the authoritative detail and the controls. Do not act on a summary alone when an action is involved — open risk, live trading, or the task queue first.
• Brief reconnect gaps are normal. If a panel looks momentarily stale, the real-time feed is likely reconnecting and polling will catch it up.

Forven is a research tool. The numbers on this dashboard — equity, leaderboard rankings, PnL — are illustrative records of a research process, not predictions of future performance, and nothing here is financial advice.

Related

• Operations & system controls — start/stop the system, execution mode, scheduler status, and system health
• Risk & safety — the gauges, circuit breakers, and kill-switch the dashboard surfaces
• The pipeline — the full strategy lifecycle the pipeline panel summarizes
• Health monitoring — how the green / amber / red component states are computed