Unfortunately an oversight in our error handling code caused the queue to be backed up for two hours. We resolved the issue, but queue processing did not pick up quickly enough, therefore we decided to clear the queue of unprocessed data.
Besides the obvious data missing from your reports, you may also have received notifications with event durations that do not match the actual duration.
We are now checking our code to see if there are any other conditions that could stop queue processing like it did today, as well as working on ways to keep the queue from backing up the way it did.
We are sorry for the inconvenience.