I keep getting AWS errors in `gcf deploy` ``` Gl...
# orm-help
t
I keep getting AWS errors in
gcf deploy
Copy code
Global
    ✖ software.amazon.awssdk.services.lambda.model.CodeStorageExceededException: Code storage limit exceeded. (Service: null; Status Code: 400; ErrorCode: CodeStorageExceededException; Request ID: 10cb8638-4bfe-11e8-a484-3739a8388b97)
@nilan this happened before and the service was listed as down on
status.graph.cool
with no deploying for almost 2 weeks…is that happening again?
😢 1
d
I had this happen to me as well about two days ago. It worked after I tried enough times.
t
@Dan yeah, the first time this error came up deployments were down, no matter how many times you tried for about 2 weeks..and they made some changes to avoid it… now it’s been happening intermittently for the past 2 days.. but the past 3 hours I haven’t been able to deploy at all after trying 20+ times.
d
@tsdexter. Yea I was with you during that 2 week downtime. It was frustrating. I’m migrating my graphcool deployment to prisma. Prisma pretty much just sits in front of your database and provides the schema. Then you have an express like server that does authentication, permissions, and functions. I.E. graphcool doesn’t manage any of that if you use prisma. The reason I say all this is because if you migrate over to prisma you can deploy to aws or google cloud and not have to worry about the reliability of prisma cloud or graphcool cloud. Prisma is just sql database and two node servers.
👍 1
t
we’re planning out a migration…but dont have the resources atm…
types.graphql
is ~900 lines.. with lots of custom resolvers… am I correct in thinking that for prisma you have to write every resolver? including
Post
,
createPost
,
updatePost
,
deletePost
for every type….or do you still get an automatic API from the schema and just write custom resolvers? We’re also looking at pushing graphcool service to kubernetes like: https://github.com/abeagley/graphcool-framework/tree/master/examples/typescript-kubernetes-cluster-with-gateway https://github.com/graphcool/prisma/pull/1324
s
Been having this issue for the last week. Brought it to Nilan a few days ago after I was about to throw my computer. Agreed with Daniel, this isn't a viable solution anymore. I need to host my own. Interested in hearing a response to Thomas as we are in the same boat. We don't have the resources to completely rewrite this. Is it even possible to use Graphcool-Framework and host it yourself?
m
Indeed, this isnt a viable solution. I've passed the entire day running a loop to try the deployment without success, Did you got any answer from maintainers? @senorcodecat
s
@marlonbrgomes Nope. Just "looking into this" which is rather absurd given it's been almost a week.
😞 1
t
@marlonbrgomes @senorcodecat deployment working again... I'm currently investigating which is more beneficial, migrate to prisma or deploy Graphcool-Framework to kubernetes on my own google cloud (which is possible, but extra work)
s
Thanks for the update, sir @tsdexter. I'm looking into the same. I believe Prisma would be far more work on our end.
t
no problem @senorcodecat. glad to see we’re not the only ones running on graphcool with all these prisma threads… let me know when you have it running in the cloud (ideally google) and I’ll pay you to set up the same for me!
s
@tsdexter I appreciate the offer, but that won't be possible haha. Launched my company last week. 😛
t
@senorcodecat haha, I hear that. I'm in the same boat. we've been running with ~5-10 enterprise clients for a few months now but are launching the public version this month. Spare dev time to migrate to another new system is hard to come by... we'll probably do prisma and apollo v2.x at the same time since they both have a lot of refactoring in the api section... Good luck with your app 👍