I am wondering if there is any way to migrate Graylog from a old server to another server with persisting all the user, stream, index set etc. configuration but discarding all the log messages. New server has new DNS names, IP:s etc., and the idea would be to have two parallel environments, so that applications can gradually migrate to the new Graylog with similar stream configuration, and after all applications have changed their configuration to send log messages to the new Graylog server, the old one could be turned off. Old server is using 2.2.3 version of Graylog, but if Graylog, Mongo and Elasticsearch versions could be updated to a newer minor release as well, that would be a plus as well.
I briefly tried just dumping mongodb and config from old Graylog, and starting it in the new server using docker-compose with proper mounts, which partially works as it loads all the users, streams etc. correctly, but Mongo seems to have references to old host names, and the Elasticsearch doesn’t work with it (might be host name configuration) when started from the scratch with empty folder. This doesn’t feel like a proper way of migrating either, as it is just a normal Mongo dump. So I am just wondering if there is any proper way of achieving what I want, or is it just easier to create everything from users to index sets by hand in the new server?