hi all, I’m have a question does not related direct with Prisma, but related to my system. Anybody please advise me. I want to build a webpage using nextjs, and api using node + prisma. My estimate system will have capacity to serve for 10k concurrent user (CCU). How to determine how many resource to use? for example. I’m using Google cloud, how many CPU, RAM necessary. Many thanks.

Hello @John Smeeth I think it depends on your application architecture. So it’s hard to say unless you can test it with traffic and adjust settings accordingly 🙂 I’m not too familiar with Google Cloud but since they are direct competitor with AWS, they may have an equivalent service to Fargate? Fargate is a service that allows auto-scaling of your containerised applications. This means you could set minimum and maximum amount of containers and their CPU / RAM. It will scale horizontally depending on the settings you choose ( e.g. at 60% CPU, spin up another container to split the load ) I normally set a minimum of 2 containers for my node+prisma backend and check the logs to see how traffic affects the services in terms of CPU / RAM and update the config from there. Hope this helps!

Many thank @Eddy Nguyen

I can definitely recommend watching