Right now FTDC reports the base value (4096) for block_size on follower nodes. This is not the same value returned by a statistics=(size) cursor, which instead uses the checkpoint size from checkpoint metadata (refreshed with each checkpoint). 

This ticket is to investigate if the other cursor types should also use the wt_block_disagg_ckpt_size fast path to record the block_size on followers, and if so to implement this change.