Why is everyone concerned with scaling databases and clustering before even launching?

DBs are great when everything works fine, but then when it stops working fine for whatever reason, your hair are on fire and you end up having to learn real quick a whole lot of stuff about your DB... then you wish:

- You had learned and practiced this stuff ahead of time; or,

- Had paid someone to do it for you, so you can focus on your core business.

Focusing prematurely on scaling and replication is the exact opposite of "focusing on your core business". The only thing that I always recommend in terms of databases is having a backup system in place. When need be and you're one of those 1% that needs scaling, aws, google and azure are there for a warm welcome.

Making sure that scaling is practical for the platform(s) you choose is not the same as focusing your efforts on premature scaling.

I said, I "hope" to grow it to scale. Not my top concern ATM.

I have a postgres instance doing 3k/qps on a single core server without braking a sweat and at 9-10% cpu. Databases are underrated.