At the bottom of the insert staging process, we need to store the index of the current measurement in the user batch into the WriteBatch for use in retries and error handling in the upper time-series write layers.

This ticket will need to add a vector to the WriteBatch, and plumb the index one layer further down in order to add it to the WriteBatch at the correct layer.

It would be handy to define an alias for the size_t in BatchedInsertTuple, to something like userBatchIndex.