Please forgive the delayed response (now that it’s been quite a few hours since the issue has been resolved), but I was traveling for most of the day.
We had a bit of a backup due to a spike in traffic around midnight in Japan. This only affected User Streams, Site Streams, and some internal stream consumers, such as TweetDeck. Some messages on the streams were merely delayed, and some had to be dropped altogether to resolve the issue.
No word yet on whether we’re be “replaying” the missing messages, but it’s unlikely.
Sorry for any inconvenience, and thanks for your patience.