Can asynq add a scheduled task dynamically using the Rset API trigger?

[daicheng123]

just like python celery,thanks

[hibiken]

I'm not familiar with Celery's API you are referring to. Would you mind providing a link or adding a detailed explanation?

[daicheng123]

@transaction.atomic()
@method_decorator(advance_logger('create a cron task by user'))
def post(self, request):
    cron_info = request.data['cron_form']
    work_id = request.data['work_id']
    work_object = Cluster_workinfo.objects.get(id=work_id)
    task_info = {
        'cluster_id': work_object.cluster_id,
        'work_id': work_object.id,
        'work_name': work_object.work_name
    }
    task_type = 1
    s_id = transaction.savepoint()
    try:
        schedule, _ = CrontabSchedule.objects.get_or_create(
                                                            minute=cron_info['min'],
                                                            hour=cron_info['hour'],
                                                            day_of_month=cron_info['day'],
                                                            month_of_year=cron_info['mon'],
                                                            day_of_week=cron_info['week'],
                                                            timezone='Asia/Shanghai'
                                                        )
        pt_object = PeriodicTask.objects.create(
                                                crontab=schedule,
                                                name=work_object.work_name,
                                                task='sysintegration.tasks.cluster_task',
                                                args=json.dumps([task_info, task_type]),
                                                enabled=cron_info['cronRun'])

just like this

[kamikazechaser]

Yes it can, you will have to write your own HTTP interface and queue the the periodic job using the built in Go client or use something like -> https://github.com/newlife/asynq-py. The worker/processor will still have to be written in Go.

Note: Afaik, this isn't a Celery like protocol that can allow you to write both client and processors in any language. See #105