Hi, we are running a High-avalability cluster using docker compose, but we seem to have a problem with the nodes finding each other. We have a nodes file containing the two names of the containers, and on initial startup they can find each other. However, if one of the servers gets restarted and therefore gets a new IP, it seems like the old ip-adresses are still being used for trying to re-connect to the cluster. Is there a way to force the typesense-server to re-lookup ip-adresses on reboot? As a note, the /data folder is mounted to external disk, so state is preserved even on reboot. Thanks in advance