Let $(M,g)$ be a Riemannian manifold, and $p\in M$. As $R>0$ increases, the topology of the ball $B(p,R)$ changes, but the changes happen only at a Lebesgue measure zero set of $R$. For instance, it can only change near $R$ when $R$ is a critical value for the distance function $d(p,\cdot)$, in the sense of Grove-Shiohama, while work of Rifford implies that a.e. $R$ is a regular value for $d(p,\cdot)$.
Question 1: Is it true that for a.e. $R$, when $\epsilon$ is small, there's a diffeomorphism between the open balls ${B(p,R)}$ and ${B(p,R+\epsilon)}$ that's $C^\infty$-close to the identity? (In particular, I don't want derivatives to blow up as you near the boundaries of the balls.)
If the boundaries of these balls were smooth codim 1 submanifolds, you should be able to do this by an easy isotopy, but I don't know how to do it in general, or if there's some meaningful piecewise-smooth structure on the boundary of (almost every) Riemannian metric ball.
Also, if the answer is yes and the proof is too easy, how about:
Question 2: Is it true that for a.e. $R$, if $g'$ is another Riemannian metric that is $C^\infty$-close to $g$, there's a diffeomorphism between the open balls ${B_{g}(p,R)}$ and ${B_{g'}(p,R)}$ that's $C^\infty$-close to the identity?
I'll accept answers to either question, obviously.