Hot keys are the primary one. They destroys your "average" calculations for your throughput. Bulk loading data is the other gotcha I've run into. Had a beauti

Hot keys are the primary one. They destroys your "average" calculations for your throughput. Bulk loading data is the other gotcha I've run into. Had a beautiful use case for steady read performance of a batch dataset that was incredibly economical on Dynamo but the cost/time for loading the dataset into Dynamo was totally prohibitive. Basically Dynamo is great for constant read/write of very small, randomly distributed documents. Once you are out of thay zone things can hey dicey fast.