Catch up of a new read-replica or a core server does not work

According to my understanding when a new read-replica or a new core server is added to a healthy cluster they should start synchronization. This behaviour is described here:

Now we are trying to add a new core server (or a replica) to a cluster but with no-downtime. Initially, we tried to add an empty node to a cluster but it didn't catch up. So now we are trying to restore a database from a backup and then try to add a node to a cluster. However, we observed that if there are any changes e.g. only adding 10 nodes then a new core server or a new replica cannot catch up with a cluster even after 30 minutes.

Do you have any idea what may be wrong?