In the tree, and in certain dialogues (simple query builder, aggregation builder, etc) dbKoda will generate a list of attributes in a collection. This list may not include all the attributes found in every collection.
We sample a relatively small number of documents (a couple of hundred ) in a collection to find attributes. If every collection has the same attributes then all will be well. However, if there is a lot of variability in the collection then we can't guarantee that all attributes will be found. You might therefore think of the list as "frequently found attributes".