Sometimes rq's job queue will loss job which cend a recursive function with `enqueue_in` to schedule.

This is my code snippet from a custom django app's command line.

if __name__ == "__main__":
    scheduler.enqueue_in(time_delta=timedelta(seconds=1), func=loop, job_id="id-test", on_failure=loop_failure)

@job("high", result_ttl=600)
def loop():
    try:
        logger.info(f"loop wakeup at  {datetime.now().isoformat()}")
        sleep = service_loop_schedule()
    except Exception as err:
        logger.error(f"loop err:{err} \n{traceback.format_exc()}")
        sleep = [1]
    logger.info(f"start_schedule after loop sleep {sleep}")

    def _requeue():
        try:
            scheduler: DjangoScheduler = django_rq.get_scheduler()
            scheduler.enqueue_in(time_delta=timedelta(seconds=min(sleep)), func=loop, job_id="id-test", on_failure=loop_failure)
        except Exception as err:
            logger.error(f"loop _requeue err:{err} \n{traceback.format_exc()}")
            return err

    start = time.time()
    failed = _requeue()
    while failed:
        time.sleep(random.randint(1, 2))
        failed = _requeue()
        if time.time()-start > 10 and failed:
            logger.critical(f"loop _requeue error: {failed}")
            break
def loop_failure(job: Job, conn: Redis, exec_info):
    logger.error(f"loop_failure job:{job.id} info:{exec_info}")

The code in django app's command will run normally for a moment, but it will loss 'loop' job at some time without any other error occur. Anyone knows the reason why?

rq / rq-scheduler

Sometimes rq's job queue will loss job which cend a recursive function with `enqueue_in` to schedule. #294