Get the observables into the ensemble results

[jlperla]

Right now the ensemble stuff works for getting the underlying dynamics (i.e., the u) but doesn't provide the z from the solutions. Might be nice to have a way to get that in the ensemble simulations as well. For example, a setting to concatenate them/etc.

[ChrisRackauckas]

Yeah, I noticed this. We need some way to make the ensembles a bit more general, for here and the optimization results.

[jlperla]

Or to be more specific for reading later, the sol.u[i].z works for i the ensemble number, but the summary statistics/etc. are trickier. So it isn't like the information is lost, just less clean for pretty access.