Any ideas on this topic? 🙌

we treat prisma as you would treat any ORM, data access layer. So you can totally reuse your prisma endpoint in as many graphql servers as you want so we might have a website with an app and users section, each one is implemented as a node app with graphqlyoga, and SSRnd JS website, and they both have their own "internal" schema, but the resolvers all talk to prisma to get their data

Right.. I was thinking in terms of traffic, that it might be a bottleneck to have one endpoint to multiple services. I'm not in DevOps so I might be underestimating Prisma's self hosted network capabilities

@Dukuo I think it's a valid concern -- one that I share. But I'm going with "premature optimization" and pressing ahead. I haven't seen real, at scale Prisma server benchmarks, but anecdotally they've built a very performant piece of software. Combine Prisma with with AWS RDS / Aurora, and if I need to deal with traffic problems, I've done a lot of other things right. 🙂

Yeah I agree in that it's very early to start figuring out facebook-level scaling 🤣 thanks a lot @Avi Block @lawjolla