Monitor
Plugin ID: gg.empostor.monitorConfig file: None — no configuration required
| Reference | Where |
|---|---|
| Configuration | None |
| Commands | None |
| Admin Panel | None |
| HTTP API | Endpoints |
API
GET /api/monitor/status
Detailed server status for dashboards.
json
{
"status": "running",
"timestamp": "2026-05-23T12:00:00.0000000Z",
"uptime_seconds": 86400,
"uptime_display": "1d 0h 0m",
"version": "1.0.0",
"runtime": ".NET 8.0.0",
"platform": "Linux 5.15.0",
"processors": 4,
"memory_mb": 256,
"game_count": 3,
"player_count": 22,
"active_connections": 30,
"games": [
{
"code": "ABCDEF",
"state": "Started",
"map": "Skeld",
"player_count": 10,
"host": "PlayerName"
}
]
}| Field | Type | Description |
|---|---|---|
status | string | Always "running" when the endpoint answers |
timestamp | string | ISO 8601 UTC timestamp |
uptime_seconds | number | Seconds since the server process started |
uptime_display | string | Human-readable uptime |
version | string | Server version |
runtime | string | .NET runtime description |
platform | string | OS description |
processors | number | CPU core count |
memory_mb | number | Working set memory in MB |
game_count | number | Total active games |
player_count | number | Players currently in games |
active_connections | number | Total active connections |
games | array | Per-game detail objects (see above) |
GET /api/monitor/health
Lightweight probe for load balancers and uptime monitors. Returns HTTP 200 when healthy.
json
{
"status": "healthy",
"timestamp": "2026-05-23T12:00:00.0000000Z"
}Use health as the probe
health returns a two-field payload and is far cheaper than status. Poll it at high frequency and reserve status for dashboards.
Related
- Player Log — persistent per-player history rather than live snapshots.
- Admin Panel — the built-in dashboard for the same data.