Docker swarm servers not reacting to down servers very fast

When a server goes down traefik stays thinking it is still up for quite a long time before removing it from the list of servers causing 502s is there any way to make it more reactive?