Dashboard Reference¶
The ArqonHPO Dashboard is a web-based UI for monitoring and interacting with running optimization jobs.

Starting the Dashboard¶
Options¶
| Flag | Default | Description |
|---|---|---|
--state | (required) | Path to solver state file |
--events | (optional) | Path to events log file |
--actions | (optional) | Path to actions log file |
--addr | 127.0.0.1:3030 | Address to bind HTTP server |
Then open http://127.0.0.1:3030 in your browser.
Interface Overview¶
The dashboard provides:
- Status Panel — Current phase, budget usage, best value
- Parameters Table — Current best parameters
- History Chart — Objective value over iterations
- Event Log — Real-time event stream
- Actions Panel — Send commands to the solver
REST API Endpoints¶
The dashboard serves a REST API for programmatic access:
GET /api/state¶
Returns the full solver state.
Response:
{
"config": {
"seed": 42,
"budget": 100,
"bounds": { ... }
},
"history": [
{"params": {"x": 0.5}, "value": 1.23, "cost": 1.0},
...
],
"phase": "Refine",
"best_idx": 42
}
GET /api/summary¶
Returns a compact summary of optimization progress.
Response:
{
"phase": "Refine",
"budget_used": 45,
"budget_total": 100,
"best_value": 0.0234,
"best_params": { "x": 1.98, "y": -0.99 },
"history_len": 45
}
GET /api/events¶
Returns recent events from the events log.
Query Parameters:
| Param | Type | Description |
|---|---|---|
since | int | Unix timestamp, return events after this time |
limit | int | Maximum events to return (default: 100) |
Response:
{
"events": [
{ "ts": 1704067201, "type": "ask", "batch": 4 },
{ "ts": 1704067202, "type": "tell", "count": 4 }
]
}
GET /api/actions¶
Returns recent actions from the actions log.
Response:
POST /api/actions¶
Send an action to the solver.
curl -X POST http://127.0.0.1:3030/api/actions \
-H "Content-Type: application/json" \
-d '{"type": "pause"}'
Action Types:
| Type | Description |
|---|---|
pause | Pause optimization |
resume | Resume optimization |
stop | Stop optimization (cannot resume) |
rollback | Revert to previous best |
Response:
Security Considerations¶
[!WARNING] The dashboard does not have authentication. Bind only to
127.0.0.1or use a reverse proxy with auth.
For production use:
Prometheus Metrics¶
The dashboard also exposes Prometheus metrics if --metrics-addr is set:
Then scrape http://127.0.0.1:9898/metrics.
Next Steps¶
- TUI Reference — Terminal-based monitoring
- Observability — Prometheus metrics
- CLI Reference — Full CLI documentation