Fix handling of iops and throughput during `manage_pool`

[rfrenchseti]

Right now get_instance_pricing uses the provisioned IOPS and throughput returned by get_available_instance_types. There needs to be a way to override this, so that when list_running_instances returns the actual provisioned IOPS and throughput for a running instance, we can use those numbers to compute the price. The most efficient way to do this would be do have some kind of update_pricing_totals function that recomputed the boot disk price using these new values.

We also need to handle the boot disks being different sizes.