{% extends "_layout.html" %} {% set title = "The run was interrupted" %} {% set status = "Not an HTTP response" %} {% set description = "Never an HTTP response. A type on a job’s or node’s error object: the environment ended the run." %} {% block main %}

The environment ended the run, and nothing in the job caused it: the scheduler lost the job’s work or a node’s, a queued job included; a node was preempted; its compute node failed; or its image could not be pulled. One interrupted node fails the whole job with this type.

A queued job that ends failed carries this type when the scheduler lost it, and run-failed when a deployment limit ended it.

It covers what scheduler-lost meant, which is retired and never raised again.

What to do

  1. Resubmit the job unchanged. It may work: the job was not the problem.
  2. If it keeps happening, quote the job’s id to whoever runs the deployment.

The body

{
  "id":    "01J9...",
  "state": "failed",
  "error": {
    "type":   "https://sc.example.com/server-errors/run-interrupted",
    "title":  "The run was interrupted",
    "detail": "..."
  }
}

No status: the type is never an HTTP response. detail is prose and may be reworded at any time. Branch on type — never parse detail.

Where it comes from

GET /v1/jobs/{id} As the job’s error.type, and as a failed node’s error.type, whose detail names an image that would not pull.

Getting help with your specific failure

This type arrives inside a job, not on a request, so there is no trace_id on it. Quote the job’s id when you escalate — it is the one identifier that finds the run on the server, its scheduler job and its logs.

This page documents the error type. It is the same page for everyone who hits it and it cannot see your job — the job id is what connects the two.

What this is not

run-failed is the run itself failing, a time or memory limit included, in a way resubmitting unchanged repeats. This one is the environment, and resubmitting may work.

{% endblock %}