Annotations of key-value regions and text heavy tables

I am a little confused. In the DocLayNet Labeling Guide you mention that a glossary should be a table (cf. 8.6.2, p.82). But at that time there was no Key-Value Region class. So would a glossary like the one shown in the guide be considered a Key-Value Region nowadays?

Also, I have trouble with a certain type of table. There is one very large table where most of the content is text, and within that text some parts are actually Key-Value Regions. How should this be handled — as just a table, or should it be annotated more precisely?