Health endpoints

Enabled with Health:Enabled, served on Health:Port (default 2455). No authentication: bind to localhost or firewall the port if that matters.

GET /health ​

Service status and database connectivity.

{
  "status": "healthy",
  "service": "trignis-service",
  "uptime": "3600s",
  "timestamp": "2026-07-28T10:00:00Z",
  "version": "1.3.3",
  "checks": {
    "database": {
      "status": "ok (all)",
      "response_time_ms": 42
    }
  }
}

status is healthy only when every configured database answers. database.status is ok (all), degraded (2/3) or failed (all).

Each call opens a real connection per unique database with a 5 second timeout, so the response is cached for Health:CacheDurationSeconds (default 120).

GET /health/connections ​

Last known state of every monitored connection, from the background prober rather than a live check.

{
  "sql:production/PrimaryDatabase": {
    "is_healthy": true,
    "last_error": null
  },
  "RabbitMQ:rabbit_main": {
    "is_healthy": false,
    "last_error": "3 consecutive failure(s)"
  }
}

Keys are sql:{environment}/{connectionKey} or {queueType}:{endpointKey}. Populated 10 seconds after startup, refreshed every HealthCheckIntervalMinutes.

Only RabbitMQ, Azure Service Bus and AWS SQS are probed. Event Hubs and Kafka endpoints do not appear.

GET /health/state ​

Sync position of every tracked object.

{
  "timestamp": "2026-07-28T10:00:00Z",
  "total_environments": 1,
  "environments": [
    {
      "name": "production",
      "object_count": 2,
      "objects": [
        {
          "object_name": "Items",
          "stored_procedure_name": "web.get_itemssync",
          "last_version": 12345,
          "last_updated": "2026-07-28T09:59:30Z"
        }
      ]
    }
  ]
}

This reads state.db, so it lists objects with stored state. An object configured but not yet initialised does not appear.

GET /health/state/

The same data for one environment. Returns 404 if no state is stored under that name.

{
  "environment": "production",
  "timestamp": "2026-07-28T10:00:00Z",
  "object_count": 2,
  "objects": []
}

GET /health/deadletters ​

Dead letter counts.

{
  "TotalCount": 12,
  "LastHourCount": 0,
  "Last24HoursCount": 3,
  "Last7DaysCount": 12,
  "MostCommonError": "Connection timeout",
  "MostCommonErrorCount": 7
}

MostCommonError covers the last 24 hours. Zeroes are returned if sinkhole.db does not exist yet.

GET / ​

Service discovery when the UI is disabled: a JSON index of the routes above. With WebHost:Enabled, it redirects to /ui instead.

Monitoring ​

/health is the endpoint to poll. status is healthy or degraded; a non-200 means the process is not answering at all.

/health/deadletters is worth alerting on. A rising LastHourCount means exports are failing now, which /health alone will not tell you, since a database can be perfectly reachable while every downstream endpoint refuses.