dependency_unavailable.

What it means.

An upstream dependency (Supabase, Redis, etc.) is failing. Orbyt is degraded but will auto-recover. Retry per error.retry_hint.after_seconds.

Retry strategy.

The envelope returns retry_strategy: "retry_after_seconds" by default for this code.

Wait the number of seconds given in error.retry_hint.after_seconds (also mirrored in the Retry-After HTTP header), then retry the identical request.

Example error envelope.

Locked RFC-001 shape. The HTTP status is 503, the x-request-id header mirrors error.request_id, and the doc_url points back to this page.

{
  "error": {
    "type": "service_unavailable",
    "code": "dependency_unavailable",
    "message": "An upstream dependency (Supabase, Redis, etc.) is failing.",
    "doc_url": "https://www.orbytlabs.ai/intelligence/api/errors/dependency_unavailable",
    "request_id": "req_0123456789abcdef",
    "retry_strategy": "retry_after_seconds",
    "retry_hint": {
      "after_seconds": 60
    }
  }
}

How to fix it.

1. Honor the Retry-After header or error.retry_hint.after_seconds before retrying. Standard HTTP retry libraries handle this automatically.
2. Add exponential backoff with jitter if you see this repeatedly under load.
3. If the error persists past several retry windows, contact support with the error.request_id.

Related codes.

Other codes in the service_unavailable family.

Keep building.