Diving in headfirst
An ebook auto-generator could be lucrative if it created the right type of content. Sadly Sudoku puzzles are not that sort of content. A quick Google search (after the fact) for "free sudoku puzzles" reveals over 133,000 pages with that exact phrase - so being able to sell the generated puzzles would be impossible. At the time I didn't think to check this, as I naively thought I must be the first person to have thought of this.
I also didn't take into consideration the complexity of creating a Sudoku generator. A solver is easy because you can just run a brute force algorithm over a puzzle that you know to be valid, and it should solve it in milliseconds (provided you coded it well enough). But a generator is another question entirely, because there are very strict rules that govern what constitutes a valid Sudoku - thus puzzle creation becomes a game of "try a random permutation, check if it's valid, repeat" (in very basic terms).
When the program was eventually done and creating puzzles (after at least 2/3 months of dev time) I created the introductory book for my Sudoku Sage series. I published it with SmashWords and in 12 months only received 1 sale (and that was from a friend). The quality of the book isn't the issue here, the problem is in fact two fold: the Sudoku market was(and is) saturated, and I didn't do any marketing for my book.
Lesson not quite learned
But I fell into the same trap, a Google search now reveals that nearly 300,000 sites use the exact phrase "poker odds calculator". Meaning the target market has massive competition. I did some actual marketing this time, SEO work, product promotion, I signed up with both JVZoo and Clickbank, creating 2 separate sales pages on two different domains. And yet for some reason despite putting the affiliate rate at 75% no one seemed to want to try to sell my program. This might have been due to the quality of my sales pages, it could have been because I was a new seller and affiliates didn't want to spend time on a non-proven product, or it might have been market saturation... I'm not sure.
The bottom line was that there were no sales. Please don't make the same mistakes as me, and instead have a quick read of commercial software product development life-cycle series for more information.