Health Check Contract

Core Health Endpoint

GET /api/v1/health (no authentication required)

{
    "status": "ok",
    "version": "0.1.0",
    "commit": "abc1234",
    "built": "2026-04-01T00:00:00Z",
    "engine": "arcan",
    "mode": "standalone",
    "sealed": false,
    "uptime": "2h30m0s",
    "checks": {
        "database":   { "status": "healthy" },
        "encryption": { "status": "healthy" },
        "plugins":    { "status": "healthy", "loaded": 3, "degraded": 0 }
    }
}

Status Values

Status	Meaning
ok	All subsystems healthy and unsealed
sealed	Encryption seal active -- unseal required
degraded	One or more subsystems unhealthy

Engine Health via Ping()

Since plugins run inside the core process (sandboxed), there is no plugin process to health-check. Instead, the core calls Ping() on each loaded engine every 30 seconds to verify connectivity to the external system (database, cloud API, etc.).

Health Monitoring Rules

• Core calls Ping() on each loaded engine every 30 seconds.
• Three consecutive Ping failures mark the engine as degraded and log a warning.
• Degraded engines still accept requests (they might recover).
• Engine health is reported in the /api/v1/health response under checks.engines.