Also can someone please 🙏🙏🙏🙏🙏 explain to me how it is expected to take a prisma architecture to production? Ive been trying to get this to work for about 3 months now. I’m exceptionally familiar now with prisma server, graphql-cli, prisma-client, and now graphqlgen… but I have a client whom I’m building their app, and I can’t figure out, nor can I find docs on how to go from a local docker-compose dev instance to actual production level architecture (like on AWS). I’ve tried using the

graphql-yoga lambda-typescript

example, the

graphql create

cli flow. I’ve even tried Apex Up and Now… but why do I have to use them, and pay for their 3rd party services? I’ve been able to get Prisma-client working up on serverless… but not working, as I then have to point to a prisma server (that currently doesn’t exist, unless I use docker)…. What am I missing? Thanks!

I have been down the same road as you in the last weeks. I have been doing ec2+docker-compose+aws rds, fargate+awsrds, "docker for aws"+awsrds and now gcloud kubernetes + cloudsql. I stay with the latter because it is almost cloud agnostic and much easier than aws cloudformation in my opinion 🙂

Thanks @yolen. Unfortunately I have to stay with AWS as my client already has an existing Postgres db. I’ve been looking into cloudformation, but you’re right, it’s not fun and I haven’t got it working yet.

it is "easy" to set up an ec2-instance with docker compose but you will have no recovery/fallback and no scaling