Hi all, does anyone know if Prisma enforces a timeout on queries, and if so, how I can change it? I’m running Prisma1 in Docker, data in MySQL

Hello, I have some trouble trying to start a prisma server in a gitlab runner. I have

Health check error

, does anyone did run a prisma server on gitlab ci ?

Hello, I have a question regarding MySQL version support please. We have an active MySQL-based Prisma v1 service in production and wanted to look at our options for MySQL versioning as v5.x starts to show its age. Does v1 support any other MySQL versions other than 5.6/5.7? I can see that v2 lists v8, but I can't find a list on the v1 documentation.

Hello Guys We use prisma 1 at where i work and we ran into an issue which we find hard to resolve. we added a new field to a model we running the yarn deploy command we had this error which prevents the field from beign added. Any help will be appreciated

ERROR: Variable "$input" got invalid value {"workspaceSlug":null,"clusterName":"default","serviceName":"new-leaders","stageName":"develop"}; Expected non-nullable type String! not to be null at value.workspaceSlug.
{
  "errors": [
    {
      "message": "Variable \"$input\" got invalid value {\"workspaceSlug\":null,\"clusterName\":\"default\",\"serviceName\":\"new-leaders\",\"stageName\":\"develop\"}; Expected non-nullable type String! not to be null at value.workspaceSlug.",
      "locations": [
        {
          "line": 2,
          "column": 17
        }
      ],
      "code": 222
    }
  ],
  "status": 400
}

@Harrison Ekpobimi you should check why

workspaceSlug

since it seems to be required for that query (

Expected non-nullable type String!

) and you are not sending it

Hi there, I currently have a a server with a container on Prisma cloud. The Heroku database that Prisma cloud uses is currently approaching the limit (9,000/10,000 rows) that the free tier allows. Currently I'm stuck as to how I would upgrade the tier that my database lives on for use by Prisma cloud. Could someone help point me in the right direction? Some context: • I don't really understand where I would run

heroku pg:copy

as discussed here. • How would I let Prisma cloud "know" to promote the database to the new one? Really appreciate your help!

Hey everyone, I’m currently trying to figure out the best way to close prisma-binding connections-- my prisma server keeps getting OOM killed and I have reason to believe that my binding connections are being orphaned due to how I have my job queue set up. Does that sound right? Or does prisma-binding actually close itself?

Do I have problems scaling in prisma1?

Hi everyone! We've recently run into an issue in production with the mongo connector for prisma:

"exception": "com.mongodb.MongoWaitQueueFullException: Too many threads are already waiting for a connection. Max number of threads (maxWaitQueueSize) of 500 has been exceeded.",

Has anyone seen this before? Are setting:

- maxIdleTimeMS=60000
- maxPoolSize=150
- waitQueueMultiple=10

sensible ways to mitigate this problem? Very grateful for any pointers if someone has faced this problem before 🙏🏾

Hi there, I don't know if I am on the right channel or if someone already has answered this question but here it goes: Currently, I am running a dev Prisma1 cloud but am getting a: ERROR: Workspace .... does not exist when I run prisma deploy If I go to the _admin URL I can still see all my records but when I login the prisma cloud my workspace is gone. Can someone please help me with this issue?

Hello, does anyone know which “version” (i mean prod or normal) of the prisma server is deployed to heroku from prisma cloud ? I need to scale by adding a second prisma-server and i would like to know what is running.

hey all, I was wondering if I could get some help with this issue I'm having with prisma1: https://github.com/prisma/prisma1/issues/5132

Hello everyone ! I tried to install prisma 1 CLI with

npm i -g prisma

as usual but the command is failing with the error

Starting from v2.0.0 publish-please can't be installed globally.

. Did I miss an unexpected deprecation ? I cant seem to find any issue about this (so I created one: https://github.com/prisma/prisma1/issues/5134)

Hi all, I've run into a small issue with the client generation and wondered if anyone might have some insight. When I generate the JS client, everything works as expected, however when I attempt to generate the Golang client, I run into the following issue: Let's say I have a type like this:

type Transaction { 
  cancelledReason: TransactionCancelledReason
}

where the

TransactionCancelledReason

is an enum defined as such:

enum TransactionCancelledReason {
  CANCELLED_FULL_REVERSAL
  CANCELLED_OTHER
}

When I run

prisma generate

with prisma 1.34.0, the resulting output in my client file is this:

type Transaction struct {
}

This is clearly missing the

cancelledReason

struct and I have the same issue for any enum field on a type. Has anyone experienced this?

Before I dig too much into what it is, has anyone tried Prisma 1 in a Google Cloud Run? It seems like a perfect setup, but I'm just curious if the prisma application is compatible so to speak with the only-run-on-request approach that it seems GCR is predicated on.

I find some references, but that seems to be either Prisma 2 and/or the Prisma client (the application).

Ok, so I am trying to run my Prisma 1 docker image (added som certificates for the database connection only, to the base 1.34 Prisma image) in a Google Compute Engine VM instance. I can get it to start, and I see in the logs that it connects to the database, and I see in the logs that it listens to port :443 (it is my understanding that when the default HTTPS rule is enabled, traffic on that external port will be passed directly to the corresponding port in the instance, which then should be 443). However, running netstat seems to suggest that there is nothing running on port 443 in the instance (and not on any other port either except the DB connections on 27017 (it is a mongo database)). Anyone tried this? What can I do to debug it? I have run the container on Heroku with pretty much the same setup, but I understand that the docker host is somewhat different between Heroku and GCE (I use the default GCE host OS).

Just to reply to myself, I set this up on the app engine instead, and it seems to be working well enough.

Hello! I have a strange problem. I run Prisma 1.34 in Google Kubernetes Engine, and have been for ~1 year. I haven't done any changes, but for some unknown reason it seems that a query feature stopped working. I have a entity called "delivery" that has a connection to entity "shipment". So one delivery can have one or more shipments. I have used a filter

shipments_none: {}

to find all deliveries that has no shipments. But this has stopped working sometime the last couple of days. So basically, even though there are deliveries without shipments, they are not returned as hits. • Prisma docker version 1.34 is set explicitly in the Kubernetes manifest. So the pod could have been rescheduled, but Prisma version should be the same. • I have not done any code/config changes during the holidays - when it stopped working • Postgres 11.5 is used as database backend Any ideas how to debug/resolve this issue would be greatly appreciated

Hello there! I have begun to encounter a really weird issue recently and I believe it is in some way related to our Prisma setup. We have been using Primsa v1.34 for the better part of a year and haven't had any issues up until this point. Pretty recently (in the last month or so) we have experienced issues with long retrieval times between Prisma and our MongoDB. Requests that used to take ~1 second are now taking ~45 seconds. Nothing has changed with our deployment, in fact, the last time we deployed the Prisma container was months ago. When I logged out the times on a specific resolver, it appears that the request to retrieve all of a collection is taking ~40 seconds which should be a much quicker operation. For example, these logs:

async function trackableMetaRequest(parent, args, context, info) {
	console.log(`-----META REQUEST START----- ${new Date(Date.now())})
    const anchorMetas = await context.prisma.anchorMetas();
	console.log(`-----META REQUEST AFTER GET ANCHORS----- ${new Date(Date.now())});

... more operations
}

End up printing:

-----META REQUEST START----- Fri Jan 22 2021 20:09:16 GMT+0000 (Coordinated Universal Time)
-----META REQUEST AFTER GET ANCHORS----- Fri Jan 22 2021 20:09:59 GMT+0000 (Coordinated Universal Time)

As you can see, there is a 43 second lull for this request which was not happening before. It used to take < 1 sec. We are hosted on AWS through an EC2 instance t2.medium tier running Docker. We have NGINX running reverse proxies between graphql and prisma as well, all through the Docker network. If anyone has any pointers or has experienced something similar, we'd appreciate the help! Thanks!

Anyone have any helpful thoughts on

ERROR: Authentication token is invalid: Token can't be decoded: Invalid signature for this token or wrong algorithm.

- I'm fully logged in and I'm able to generate a token but I can't deploy

Hey folks, we have a CPU leak somewhere on prisma@1.34 We start a big 8 core instance with autoscaling up to 4 instances. Prisma starts to slowly eat all 8 cores after being up for a while, autoscales to 32 cores and eventually just dies. The

LOG_LEVEL

is set to

DEBUG

but i can’t really see anything meaningful to me. Logs consist of the same loop over and over:

: { 
:  ...
:  /* SCHEMA */
: }
: NodeSelector(com.prisma.shared.models.Model@76c53d76,ScalarField(id,com.prisma.shared.models.TypeIdentifier$Cuid$@4b522eb2,true,false,false,false,None,None,None,FieldTemplate(id,com.prisma.shared.models.TypeIdentifier$Cuid$@4b522eb2,true,false,false,false,false,None,None,None,None,None,Some(IdBehaviour(com.prisma.shared.models.FieldBehaviour$IdStrategy$Auto$@2984dc30,None))),com.prisma.shared.models.Model@76c53d76),StringIdGCValue(ckkonhbow01af0889jmoa6apq)),PrismaArgs(RootGCValue(Map(impressionsCount -> IntGCValue(100), updatedAt -> DateTimeGCValue(2021-02-04T15:20:09.114Z)))),Vector(),Vector(),Vector(),Vector(),Vector(),Vector(),Vector(),Vector(),Vector(),Vector()))))
: #11 Commit (usually some sequence number there)
A bunch of FieldTemplate(...) and RelationTemplate(...) lines

If anyone here have some insight into how can we get down to the cpu leak I would really appreciate it 🙂

@here Has anyone performed a Prisma 1 to Prisma 2 migration? Who can contract with our team to help with the migration?

@Artur You'll probably find those who migrated in #general 🙂

Hey all. I'm wondering if anybody has implemented a version of Prisma1 where they have extended it to support graphql unions. I have a usecase for them, but am having some trouble bending the prisma client to fit my requirements

There is a sudden spike in CPU usage with Prisma 1. I have 4 CPUs and 16 GB of RAM. How could I know what is causing the spike in CPU usage of Prisma 1? All 4 CPUs are at 100%

Hi guys, I’m trying to make use of connection pooling with pg_bouncer using a managed postgres database on Digital Ocean. When the Prisma 1 server tries to connect the following error is thrown:

Exception in thread "main" org.postgresql.util.PSQLException: ERROR: unsupported startup parameter: search_path

Any thoughts? I have tried transaction and session mode so far without any luck.

Does anyone know why the servers are hanging ?

I cant seem to connect to the prisma cloud instances