Skip to main content

Documentation Index

Fetch the complete documentation index at: https://supahooks.ar27111994.dev/llms.txt

Use this file to discover all available pages before exploring further.

The server exposes two probe endpoints for container orchestrators, load balancers, and uptime monitors. Both are rate-limited but never auth-protected, even when authKey is configured. This allows infrastructure tooling to probe the service without credentials.

​
GET /health

Liveness probe. Returns 200 OK when the server process is running. Use this to detect whether the process needs to be restarted.

​
Response example

{
  "status": "healthy",
  "uptime": 3600,
  "timestamp": "2026-01-30T12:00:00.000Z",
  "memory": {
    "heapUsed": 50,
    "heapTotal": 100,
    "rss": 140,
    "unit": "MB"
  }
}
​
status
string
Always healthy when the process is responsive.
​
uptime
number
Process uptime in seconds.
​
timestamp
string
ISO 8601 timestamp of the response.
​
memory
object
Current Node.js memory usage. All values are in megabytes (unit: "MB").

​
GET /ready

Readiness probe. Returns 200 OK when the service is ready to handle traffic — specifically, when DuckDB and the webhook manager are initialized. Use this to hold traffic away from an instance that is still starting up or is mid-shutdown.

​
Response example (ready)

{
  "status": "ready",
  "timestamp": "2026-01-30T12:00:00.000Z",
  "checks": {
    "database": {
      "status": "ok"
    },
    "webhooks": {
      "status": "ok",
      "message": "1 active webhook(s)"
    }
  }
}

​
Response example (not ready — 503)

When any dependency check fails, the route returns 503 Service Unavailable and changes the top-level status to not_ready. The checks map shows which subsystem failed. 
{
  "status": "not_ready",
  "timestamp": "2026-01-30T12:00:00.000Z",
  "checks": {
    "database": {
      "status": "error",
      "message": "DuckDB not initialized"
    },
    "webhooks": {
      "status": "ok",
      "message": "1 active webhook(s)"
    }
  }
}
​
status
string
ready when all checks pass; not_ready when any check fails.
​
checks
object
Per-subsystem check results. Each entry has a status of ok or error and an optional message.

​
Graceful shutdown behavior

During intentional shutdown, the HTTP listener is drained before internal services wind down. This means probe traffic never reaches a partially torn-down state — the server stops accepting new connections cleanly before shutting down.