- around 500-1K ads will be published per day. Login required
- around 10K-20K visitors per day. No login required
- simple search (by category, free text field) and sorting (by date). No login required
- one paginated homepage where ads are displayed by date, and one page to see the details of a specific ad
- no images. It's all text-based
My plan is to use nginx + Go + SQLite + redis. Big assumption I'm making here is that the website is heavy on reads and very lightweight on writes (reasonable assumption since I maintained a similar website some years ago, so I have some experience in this regard). Also, I don't see the website growing in the near future (I have numbers that say that traffic will remain as stated above at least for 1 year).
Do you think SQLite will be enough? Main reason to not use MySQL/Postgres is: maintenance, simplicity but above all I think it will be overkill. Also, I'm planning to heavily cache searches (either on nginx side or redis). So, what do you think?
Plus, if you build it in a way that enables you to scale up (vertically to a 'better' database or horizontally to more SQLite databases) it won't even matter.
Straight fro. The horse's mouth
https://www.sqlite.org/whentouse.html
Below 100k hits should be fine. Note its hits not users. But, have used Sqlite on 25k visitors per day site with no issues.