SQLite Might *Actually* Be All You Need
Dec 02, 2026 10:00 AM Pacific
Members join meetups live and can watch the recordings
Join the community to take part in this meetup.
This talk dives into running SQLite in production and why, surprisingly, it might be one of the best production DBs you can choose.
Here's something I haven't really talked about yet: Talk Python doesn't run on MongoDB anymore. After years on Mongo, I moved most of our sites to SQLite, the database we file under "fine for prototypes and phone apps." I knew an in-process database had some speed advantages. I didn't expect it to be 40 to 300 times faster.
Now, looking up a guest by id takes 1.5 microseconds, over 600,000 calls a second from plain Python. A single-row write commits in about 10 microseconds. Against MongoDB and PostgreSQL over a 10 Gb LAN, the same queries ran that much slower, mostly because every call pays a network round trip that SQLite never makes. I'll show the real-world benchmarks from our web apps, not just your standard "no-op" style examples.
But then there's the question that scares people off: what happens when the server dies? This is where the big surprises live. I'll introduce you to Litestream. Litestream streams every change to S3-compatible storage and gives us 72 hours of point-in-time recovery across eight production databases. I'd bet that beats the backups behind most apps running today. And in about 10 seconds, I can restore production to my laptop, in sync with the live site.