{% extends "_layout.html" %} {% set title = "The server could not get this job ready" %} {% set status = "Not an HTTP response" %} {% set description = "Never an HTTP response. A type on a job’s error object." %} {% block main %}

The server could not get the job ready to queue, for its own reasons, after retrying: a dispatch the scheduler refused; an index or a registry that did not answer; a build of a node’s Python environment that was lost; an index that needs a credential the server cannot supply; an image the manifest is read in that could not be pulled; or a store that did not answer while copying in the results of an earlier job. Nothing you sent caused it, and the job ends failed. detail says which.

What to do

  1. Submit the job again later: the server already retried for as long as it waits.
  2. If detail names an index that needs a credential, that is for the operator to set up, not for you to change.

The body

{
  "id":    "01J9...",
  "state": "failed",
  "error": {
    "type":   "https://sc.example.com/server-errors/staging-failed",
    "title":  "The server could not get this job ready",
    "detail": "..."
  }
}

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 the reason on its staging → failed transition. Never an HTTP response: staging runs after the submit’s 202, so there is no request left to answer.

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 and what it tried.

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

software-unavailable with reason: "uninstallable" is a Python package that will not install for the node’s image: the job is rejected, and the fix is yours. This one is the server’s, and the job is failed.

run-interrupted is a job the scheduler held and the environment ended. This one never reached it.

staging-timed-out is a job that ran past its own max_staging_seconds while it staged, a build that ran out of time among them. That one is the job’s limit, and resubmitting it unchanged times out again; this one is the server’s.

{% endblock %}