BucketCatalog::abort takes an optional Status which it uses as the error code to abort all batches on the same bucket. If no Status is provided, then it defaults to using TimeseriesBucketCleared. However, it would be better to require callers to specify an abort status. This way we won't potentially lose the error that caused the abort across batches, by converting to TimeseriesBucketCleared when really it was a different status that was the cause.