GET/healthfree, no auth

Health check

Component-level status for the gateway and its dependencies. No bearer token required, no units consumed. Use it for uptime probes and as the first call in a multi-step job to short-circuit before you spend.

Request

curl https://data.leadleap.net/health

Healthy response

{
  "status": "ok",
  "service": "leadleap-mcp",
  "components": {
    "lake_coordinator": { "status": "ok" },
    "customer_db":      { "status": "ok", "pool_size": 1, "pool_available": 0 },
    "redis":            { "status": "ok", "used_memory_human": "2.02M" }
  }
}

Degraded response

Top-level status becomes degraded when any component is unreachable. Individual components carry their own status with an error string when applicable. The gateway still answers requests when degraded - federated /v1/find may return partial results with a warning.

{
  "status": "degraded",
  "service": "leadleap-mcp",
  "components": {
    "lake_coordinator": { "status": "ok" },
    "customer_db":      { "status": "unreachable", "error": "..." },
    "redis":            { "status": "ok", "used_memory_human": "2.02M" }
  }
}

Components

lake_coordinator - the data lake fan-out service. If down, /v1/find and /v1/sources return errors.
customer_db - the metadata Postgres. If down, key authentication still works (it's cached) but quota writes degrade gracefully.
redis - quota state. If down, the gateway fails open on rate limits to avoid blocking customers due to internal infra issues.

Free for a reason - point uptime monitors at it without worrying about quota. See Rate limits for the behaviour when one of these components reports degraded.