I’m exploring event sourcing as a pattern and wondering how it fits into a cloud database architecture. How do you structure the event log, and what kind of storage or databases are best for it? Also, how do you deal with replays or data corrections?