Ask HN: I think most SaaS can work well SQLite. Prove me wrong

14 points by sh_tomer ↗ HN
I think most SaaS vendors over-complicate things, especially in the beginning of the startup journey. Using simple technologies like SQLite can significantly reduce the complexity of the project, when compared to other database technologies.

what do you think?

18 comments

[ 5.0 ms ] story [ 56.4 ms ] thread
Yes, especially if they are very read heavy. Probably wouldn't go with sqlite for say a analytics saas app. And I probably would think about wether there are some functions in for example postgresql that aren't as good in sqlite. But yea most saas can probably run on sqlite.
If you slice per separate physical database file per customer and your clients aren't big corporations that might have lots of concurrent writes.
Using managed technologies such as PostgreSQL will be simpler and often cheaper than SQLite, if you'll have to manage it yourself, incl. scaling, backups, updates.
You'd be hard pressed to find a less controversial opinion for HN.
More tooling on Postgres. Why would you use SQLite? It’s great for embedded applications but if yours is networked why cobble together some networked SQLite when you can just use Postgres or another database built for that purpose?
His point is that you probably won't need networked sqlite
Post a contact email? I'd love to reach out about this exact topic!
Probably ! I use SQLIte for most of my web services, I have yet to hit a point where I need to scale them out
My SaaS business runs on SQLite, and it has been working like a champ. Backups are dead simple, and I don't have to worry about uptime, configs, network latency, and all that.

Postgres is great, and I used it in my previous business, but sometimes, it's an overkill.

Do you run it on a cloud provider, if so how do you store it, ie.: mounted volume?
I rent a dedicated server on Hetzner with additional storage volumes and back it up to Tarsnap (with other files).
It's like a repeat of c++ or python. If you can find folks & tooling for what you want to accomplish then go for it. I don't think founders who actually have paying customers concern themselves with what works best per cent of compute spend rather what allows them to improve the product, while as an outsider it may seem that they can reduce costs by choosing X over Y.
Curious if anyone has actually hit a scale where SQLite cannot handle their workload? What do you actually do at the point?

As a stupid example, lets say you’ve reached a write bottleneck.

You upgrade to a beefy machine with tons of CPU and enough memory to hold the database in RAM. This buys you enough time to migrate to something else.
(comment deleted)
IMHO, I agree with you, most businesses would probably be happy with SQLite; SQLite in WAL mode is really fast, the file size limit is basically in Terabytes and if you reach a bottleneck you probably reach a point where you can invest in more.

SQLite is quite powerful and as you said, really simple. Also, the documentation is easy to follow.