Skip to main content
Two auth-free probes for load balancers, Kubernetes, and uptime monitors. Both live outside the versioned API (/api/health, /api/ready, not /api/v1/*) and are the only routes that skip tenant + RBAC auth. Each is per-IP rate limited to 100 requests / minute (returns 429 with a Retry-After header when exceeded).

​
GET /api/health

Liveness check. Probes the database, the Redis/cache layer, and the circuit breakers. Status logic:
  • Database down → status: "degraded", HTTP 503.
  • Cache down or any circuit breaker open → status: "degraded", HTTP 200 (serving, degraded).
  • Otherwise → status: "ok", HTTP 200.

​
Response 200

{
  "status": "ok",
  "database": "ok",
  "uptime": 84213.4,
  "timestamp": "2026-07-03T12:00:00.000Z"
}
When the database probe fails the same shape is returned with status: "degraded", database: "error", and HTTP 503. uptime is the process uptime in seconds.

​
GET /api/ready

Readiness probe. Checks database and cache connectivity and reports each dependency individually.
  • Database disconnected → not ready (503).
  • Cache disconnected (ping returned falsy) → not ready (503).
  • Cache unavailable (probe threw) → not fatal; readiness still passes on the database alone.

​
Response 200

{
  "status": "ready",
  "checks": {
    "database": "connected",
    "cache": "connected"
  },
  "timestamp": "2026-07-03T12:00:00.000Z"
}
When a required dependency is down the response is status: "degraded" with HTTP 503 and the failing check marked "disconnected" (or "unavailable" if the cache probe threw).