>Do I understand correctly that the timestamps cache will be invalidated whether query is cached or not?

Yes, obviously every mutation needs to invalidate certain query cache regions, which works through the timestamps cache.

I guess that your default cache configuration might play a role here. I don’t know much about EHCache, but maybe answers like this can help: https://stackoverflow.com/questions/67993442/hhh90001006-missing-cachedefault-update-timestamps-region-was-created-on-the