Currently, PipelineScheduleWorker is fired in a short period on
our production server. We can stop this behavior by locking the
execution thread with in_lock method.