I'm sure you can extract it from the proof, but does anyone know of a reference where the radius of convergence (in terms of radius of convergence of the initial data and PDE) of the solution given by the Cauchy-Kovalevskaya Theorem is written down?

On a related, but more speculative note, I'm curious if there are any results along the following lines:  Suppose one is given a problem solvable by the Cauchy-Kovalevskaya Theorem along with a real analytic solution with uniform lower bound on its radius of convergence (alternatively which exists in some uniform strip about the hypersurface on which initial data is prescribed).  If one approximates the data of the given solution (the right notion is part of the question) by some sequence of real analytic data, can one say anything about the radius of convergence of the corresponding solutions (or size of strip the solutions exist on).