Skip to main content

service_unavailable

The instance that received your request is shutting down and stopped accepting new work. Nothing was processed.

Status: 503, with Retry-After: 5.

{
  "status": 503,
  "type": "service_unavailable",
  "message": "The service is shutting down. Retry shortly.",
  "request_id": "req_01J8K2P...",
  "doc_url": "https://docs.epostix.com/errors/service_unavailable"
}

Why you got this​

An instance is draining, almost always during a deploy. Requests already in flight are allowed to finish; new ones are turned away so that nothing is cut off mid-request.

This is a normal part of a rolling release rather than an outage, and it is brief.

How to fix it​

Wait out Retry-After and send the request again. Other instances are serving normally, so a retry will land on a healthy one.

Treat it exactly like internal_error in your client: back off, retry, cap the attempts.

Is it safe to retry​

Yes. The request was refused before any work started, so a retry cannot duplicate anything.

Reuse the same Idempotency-Key anyway if the request had one. It costs nothing and keeps your retry path identical across every kind of server-side failure.