IPs stuck when resolving service using docker network

Hi everyone,
I have some issues with a docker network in a docker swarm platform.
Right now we are using a docker network to let some APIs do request to each other.
The issue we had here is that when resolving the name of the APIs we receive more IPs than replicas of the service and even deleting the service, the “extra IPs” are still there. The only way I could solve it was to reboot all nodes. I would like to know how to prevent this issue so I don’t need to reboot all the nodes for this. It seems something like it is caching old IPs or something similar.

Additional info:
The network is an overlay network in a swarm scope.
Only the rebooted nodes get the correct ips. So I can have 2 instances of the same service resolving diferent IPs for another service

Thank you for your time, ask for any necessary information for solving this issue