Comment by xg15

Not sure I read that formula right, but isn't the data loss limited to the last 100ms (worst case)?

I don't see many single user/desktop application use cases where that kind of time range would be risky.

Maybe if you used this as the main production database with thousands of concurrent sessions. But that doesn't seem to be their main use case, is it?