I recently got a project for scaling multiple wordpress sites for around 5k req/s

The biggest mistake I made is that even though it's wordpress I accepted it, fml!

Even aws load balancer gave up.

Consider moving to .net, our services are crunching 1.2M/s, and we're middle of the pack.

Wordpress will shit it's self at the DBS side, nginx should suffice or scale the web instances if needed, but ideally scale in db replication and don't - for the love of god *unpacks a new keyboard* - have the same EC2 instance host both the web and dbs.

I agree, local db is much faster compared to rds or db on same network.

But as said, it's multiple wordpresses 😂
I still hope, nginx can pull me out of this shit.

Or using WP as offline CMS which would make the actual online website static, yielding both better performance and security.

@Fast-Nop
That's just a time/space trade-off 😁