orphaned job

[mgk]

Running without RQ Scheduler and multiple RetryWorkers I saw a job dequeued and never processed. The job was eventually orphaned. The orphaned job (not in any q, not in any registry) looked like this:

{
    "enqueued_at": "2015-08-12T16:50:02Z",
    "created_at": "2015-08-12T16:50:02Z",
    "description": "some.pkg.upload_job('some-url')",
    "origin": "upload_job",
    "status": "queued",
    "timeout": "28800"
}

The values above are all expected from the app in question. No errors were thrown and no worker processes were hard killed. There were 8 worker processes at the time so one or more might have died.

Need to understand and close or at least shrink window for job loss.