no bug at all !
This is the normal way to proceed with Web screaper.
First it deals with pagination, it starts with page 1, page 2 ... until the last page.
During this first step, no data is captured and if during this first step you stop the process, you will not recover any data.
When all the paging is done (and it can take a long time !) and if your sitemap is properly built, Step 2 begins and this is the real time when scratching is done. This scratching is done in reverse: first the last page, then page n-1, then n-2 to page 1
You can not avoid step 1. So when tuning your work, it is advisable to test your scratching with a few pages, say less than 5 pages.