I am really surprised by this, I was convinced that the table-wide capacity is re-distributed evenly across shards. Looks like you're entirely correct: http://

I am really surprised by this, I was convinced that the table-wide capacity is re-distributed evenly across shards. Looks like you're entirely correct: http://docs.aws.amazon.com/amazondynamodb/latest/developerguide/GuidelinesForTables.html#GuidelinesForTables.Partitions.Throughput http://docs.aws.amazon.com/amazondynamodb/latest/developergu... I wonder how many times I scrolled past the explanation without internalizing it.