The main takeaway is that 1 billion rows is just not that much nowadays.
A corollary is 'postgres is really all you need' unless you have irrefutable proof it isn't.
A corollary is 'postgres is really all you need' unless you have irrefutable proof it isn't.
Many businesses think "we have a lot of products". Nope, even when you have more than a million products we can load all of them into RAM, even with 4 KB of memory allocate to each product.
So a million products is small business from a "load into RAM" perspective. Easily handled on a single server.
Run a Wireshark capture while opening slack, those ~30ms round trips add up
How do you write code this badly.
I don't think the impact of org and dev team structure on code performance and quality has really been studied.
Especially if you need to have supplementary indexes to support common queries on that data.