The API takes a snapshot of the paginated set of articles at the time of the first request. Paging through articles goes through this fixed set of items. The snapshotted set becomes obsolete when a request has not been made for 2 minutes.
When you filter by
filter[url], the system always returns restricted content. Therefore, filtering by
filter[include_restricted]is not useful in these cases.