This issue is related with this line:
https://github.com/DSpace/DSpace/blob/main/dspace-api/src/main/java/org/dspace/content/CollectionServiceImpl.java#L1122

DiscoverResult discoverResult = searchService.search(context, discoverQuery);

because we didn't specify a number of expected rows. So the results with the default number of rows (I seems to be 10) are returned.

Two possible approaches could be addressed. The easy one is to define an high number for the returned rows (but this problem will again occur if that number exceeds). The other one is to loop every page result. Perhaps we can define a high number and do pagination so we don't have to do several or many solr queries.

@paulo-graca : I'd like to note a third option is to make that method support pagination similar to findCollectionsWithSubmit() here: https://github.com/DSpace/DSpace/blob/main/dspace-api/src/main/java/org/dspace/content/CollectionServiceImpl.java#L1085-L1099

Essentially, any methods which can return a large number of results ideally should support pagination parameters. That way the calling code can paginate through the results (or grab all results if it needs it). In this scenario, that'd be better than an arbitrary large number...plus it could allow for better scalability as you may not need/want to load 1000 Collection objects into memory at one time even if you have 1000 Collection objects that match an entity type.

In any case, I agree this is definitely a bug. Pulled this over to the 7.6.x maintenance board & keeping assigned to you (thanks for volunteering!)