To split or not to split up data between buckets

When would it make sense to put different document types into separate buckets? Does the presence of different document types in one bucket have any effect on index performance or performance in general?

If I have for example social network post documents and log documents, is there any advantage of putting them in different buckets?

Of course if one document type is needed to be backed up frequently and another document type is not, it would make sense I guess?

Kind of a general question, every input is welcome :slight_smile: