> if I had the option to use PostgreSQL for something I'd never ever get close to choosing SQLite over it
I had the option for a recent project. It's a niche forum-like application with around 2,000 users. Went with a monolithic design and vertical scaling (if needed), so SQLite was perfect. Every dynamic HTML page renders in under 1ms. Litestream for live-replication to a couple S3 buckets.
Running PostgreSQL for something like this would be a pain in the ass, and add a minimum 10ms of latency to every request. There would genuinely be more maintenance required for the PostgreSQL server/daemon than the application itself. It just makes no sense.
> There would genuinely be more maintenance required for the PostgreSQL server/daemon than the application itself. It just makes no sense.
I think this is the key point the original post was missing. When you’re working on a project with a DB, you should factor in the cost/overhead of the DB as well. Postgres is a wonderful DB. But it is also big and requires work to keep it running. For many projects, the extra overhead is well worth it.
(And if you’re using a “cloud” DB, the overhead is still there, but you’re explicitly paying for the privilege of making it someone else’s problem. )
But for smaller projects, or ones with less DB requirements, something small like SQLite is much more appropriate. And it nearly removes the DB overhead from the maintenance equation.
And the value of maintenance-free differs by environment. A lot of my projects are one-offs where I'm not getting paid to maintain a database system, so I don't want to maintain a database system.
Is that 10ms of network latency? I remember Postgres being pretty low latency if you run it on the same host, although the point about maintenance makes complete sense.
I had the option for a recent project. It's a niche forum-like application with around 2,000 users. Went with a monolithic design and vertical scaling (if needed), so SQLite was perfect. Every dynamic HTML page renders in under 1ms. Litestream for live-replication to a couple S3 buckets.
Running PostgreSQL for something like this would be a pain in the ass, and add a minimum 10ms of latency to every request. There would genuinely be more maintenance required for the PostgreSQL server/daemon than the application itself. It just makes no sense.