I'm assuming the limitation of 10K per month has more to do with their small cloud infrastructure budget than the actual database. Saying that, I can agree that intutively, a webserver can have slower performance than a traditional database but it would be interesting to see some benchmarks.