Is there any support for nested queries? Say I wanted to do something like getting all posts from my own friends:

posts {
  id
  author {
    id
    friends (
      where: {
        id: myID
      }
    ) {
      id
    }
  }
}

Dumb question: if running against hosted/managed prisma servers does

prisma deploy

also push resolver code to run server side or just manage the DB connections? I'm trying to learn prisma basics before porting a prototype from graph.cool over to prisma.

does anyone know of a really good up-to-date (e.g. prisma 1.8.3) tutorial on prisma & yoga? This tutorial is great, but already obsolete: https://www.howtographql.com/graphql-js/5-connecting-server-and-database/

Has GraphCool been abandoned? The name change to prisma, the lack of updates, the use of legacy in certain areas, Is graphcool effectively dead? No more updates, no more bug fixes?

Where should I start looking to resolve this error:

Cannot return null for non-nullable field Query.myQuery

?

Trying to understand how we would: A) deploy prisma to ec2 (docker is perfectly suitable here), that talks to an aws RDS database B) Deploy the graphql-yoga based api that our frontend app will consume also to ec2 (again docker is perfectly fine) — ec2 not lambda as we really need subscriptions to work. Documentation in this area seems to be sorely lacking. Fargate is interesting, but we had problems with it in our environment as we’ve already got many vpcs in play and while it didn’t complain about vpc’s during the cloudformation build, deploys simply timeout, presumably lack of communication between various pieces.

Is there any clear up to date information on how this can be achieved?

We’ve wasted most of the week screwing around with this.. not devops — we’re developers, really don’t want to have to figure out how to assemble the whole stack. Getting pretty frustrated. 😕

anyone mind sharing what a typical response time looks like for them? I’ve been testing out Prisma, first in a sandbox, now with my own cluster running in AWS via Fargate. My GQL app is deployed via up as a lambda function in the same region. Typical response times for queries seem to be between 400-600ms - not terrible, but not fantastic either (Our current GQL app with essentially the same schema is <100ms usually). Biggest surprise to me was that I didn’t see much of an improvement when switching from the sandbox to my own server. Wondering if this is just standard, or if there’s some optimization path I’m missing somewhere. Thanks!

when following the tutorial to build a graphql server with graphql-yoga, i understand it's adding a layer that defines what operations are exposed in the API to the client, which require tons of time to sit down and write out each operation that should be exposed for each type when what i want is actually to expose all of it first and then later think about what to remove. is there an easy way to do this?

I have an extremely simple mutation that works in the Playground but not in my Apollo Client. 😢

Hi, i have been toying with GraphQL for a bit now and struggling to set up a nice CRUD Api for my data without writing too much code. I really like the way Prisma is built and how its API is structured. However, we have an existing Laravel backend which has quite extensive functionality around how data is saved into the database - Prisma itself would not be able to fit our requirements. Is it possible/viable to add in a custom ApiConnector for Prisma and link that up with our existing backend?

Hello - confused about Prisma. From what I understand I have to create a SQL db somewhere and then connect it to Prisma once my Prisma Layer has been lunched on a service such as AWS Lambda. Graphcool gave me a DB instance I could start pushing to. I dont want to use, let alone setup an SQL DB! I get that people will be able to migrate from because of this - but Graphcool let me go forward.

balls. just frustrated.

@spawnia if you have an existing backend, you likely just need Yoga and then you code the resolvers yourself

Hey guys, I use graphcool on my server with docker image but my pwa is on https and my endpoint on http then i can't access to my endpoint ... An idea ?

For those with more AWS experience, what is the difference between ECS and Fargate? I know there is a tutorial for deploying a Prisma server on the latter.

ECS = Kubernetes Fargate = Scheduling engine for containers to be lamdba like So you can have ECS + Fargate and ECS + EC2 In the future you will also have EKS (Managed Kubernetes for aws) + EC2 or EKS + Fargate

ahhh okay.

The main diference is that ECS in fargate "mode" you will be payingg for compute slots like you do in lambda or you can use in EC2 "mode" and setup how many nodes you want

It's a higher abstraction so you don't need to setup cluster, nodes, autoscaling

Okay, I appreciate the clarification. I've been hunting for a guide to deploying Prisma on EC2 (or I suppose ECS+EC2).

Also: wasn't there a guide somewhere for deploying with Zeit Now?

I’m trying to use the prisma binding “exists”. I have a pretty simple request “prismaDb.exists.User({ where: { id: ‘cjh0q7xpi8fsd0b792ziqrfg1’ } });”

but I’m getting this error: “UnhandledPromiseRejectionWarning: Error: Variable “$_where” got invalid value {“where”{“id”“cjh0q7xpi8fsd0b792ziqrfg1"}}; Field “where” is not defined by type UserWhereInput.”

anyone have any idea?

yeah getting errors right now

Greetings, forgive me but this may be a GraphQL question, but I'm wondering if it's possible to do geo spatial queries in prisma. For example if I have a type of Restaurant, and a user does a query for restaurants within a 5 mile radius is it possible to set this up in prisma ?

I deployed my Prisma server with Zeit Now and am getting the following error:

request to <https://localhost:4466/> failed, reason: connect ECONNREFUSED 127.0.0.1:4466

Any thoughts?

hey folks, how can we dump a bulk of data into prisma database