Datastore (the database technology that powers Google App Engine) has a number of great features.
Although datastore is perfect as a Key/Value store that allows transactions and limited queries, it does have its limitations.
One effect of these limits is that it can difficult to do things like aggregation on tables. You should use the proper tool for the desired outcome.
Datastore is structured (typed) data. Everything that goes into Elasticsearch must be JSON data. Structured data can be fairly easily transformed into JSON data. You see where I am going with this?
This extension to gaend does successfully convert any Datastore data into Elasticsearch. You should look at your table in Datastore and perhaps consider whether you need to submit a custom mapping for Elasticsearch. If unspecified, Elasticsearch will generate a mapping for you based on the submitted JSON document. For tables made up of "primitive" data, the dynamically generated mapping is usually sufficient. Still, know that you can override the dynamic in the event that you have complex, deeply nested, or large types of data.