The former isn't the sense that I have in mind in saying that faith in god might be rational. If it were, it'd be easy to knock down faith. After all, it's irrational to have faith (epistemically) in things that don't make any sense. The concept of god depends on the concept of the supernatural. And the concept of the supernatural doesn't make sense. So, it's irrational to have faith in god.
Beyond just that problem, the commending sense of "rational" doesn't easily apply to something as difficult to specify, and to prove, as god or the supernatural generally. We usually reserve rational-as-commendation for things that make sense to do or to think. Things like financially planning for the future, or looking both ways before crossing the street, or not eating yellow snow. We commend those things by calling them "rational" because they antecedently make sense. When it comes to religious matters, though, there is real controversy about whether they actually do make sense. So, you don't have that before-the-fact assurance that religious faith makes sense, and that puts up a pretty significant roadblock to attempts to treat faith as rational in this way.
Generally, then, I think it's fairer to think of faith as rational in the second sense. That's certainly the sense that Kant is aiming at. It's rational because it forms part of human reason, not because it necessarily deserves any sort of commendation. And that does at least save some idea of faith from destruction -- having faith is believing in things that are beyond what's knowable.
However, people who talk about faith -- religious people, that is -- do actually want to commend faith, even recommend faith to others. So, they're probably not going to be satisfied with an account which holds that faith is rational descriptively. They might then start to think that I've done something wrong in thinking of faith as rational in the descriptive sense.