System

Health checks, readiness, and operational endpoints.

GET /v1/health#

Liveness check. Always returns 200 if the API process is running — doesn't inspect dependencies.

bash

1	`curl https://scaivault.scailabs.ai/v1/health`

Response:

json
{
  "status": "healthy",
  "timestamp": "2026-04-23T..."
}

No authentication. Use for load balancer health checks.

GET /v1/health/ready#

Readiness check. 200 only if ScaiVault can serve traffic (DB, Redis, encryption all OK).

Response (200 OK):

json
{
  "status": "ready",
  "checks": {
    "database": "ok",
    "redis": "ok",
    "encryption": "ok"
  }
}

Response (503):

json
{
  "status": "not_ready",
  "checks": {
    "database": "ok",
    "redis": "error",
    "encryption": "ok"
  },
  "errors": {"redis": "Connection refused"}
}

No authentication. Use for Kubernetes readiness probes.

GET /v1/health/detailed#

Full component status with latencies and resource stats.

Response:

json
{
  "status": "healthy",
  "version": "1.0.0",
  "uptime_seconds": 86400,
  "checks": {
    "database": {
      "status": "ok",
      "latency_ms": 2,
      "pool_size": 20,
      "pool_available": 18
    },
    "redis": {
      "status": "ok",
      "latency_ms": 1,
      "memory_used_mb": 128
    },
    "encryption": {
      "status": "ok",
      "key_status": "active",
      "last_rotation": "..."
    },
    "s3": {
      "status": "ok",
      "latency_ms": 15
    }
  }
}

Scope: admin.

GET /v1/metrics#

Prometheus-format metrics. Served on the main port or a separate metrics port depending on deployment config.

Metrics include request counts/latencies per endpoint class, active leases, cache hit rates, rotation queue depth, KMS call counts.

Authentication: configurable. By default, unauthenticated on internal-only port.

GET /v1/version#

Returns version info. No auth.

json
{
  "version": "1.0.0",
  "build": "abc123",
  "git_sha": "def456",
  "built_at": "2026-04-01T00:00:00Z"
}