seems like redis is crashing over there and therefore the queue is getting drowned in jobs. cheri is working on it