@tsdexter would you mind elaborating on why the prisma migration is a big feat? we're heavily considering that given some issues

@tsdexter How have you found the support, being on the pay tier? Myself, I needed quite a bit of handholding when I started a year ago and they were great. Once I got going, I've been happily developing for a year with no problems. I had a question 4 weeks ago which has gone unanswered. Now this and it seems that the support is way worse than it was before. For free tier, I'm fine with it. But I would expect better with paid tier.

are there any guides on self hosting out there?

mostly cause our schema is ~1100 lines, 25 custom resolvers, a stripe integration… we have to write everything for the prisma app layer, rebuuld permissions back in and then update every front end API query to the new format (no ‘all’, ie:

allItems

->

items

,

filter

to

where

and unique types to the

where

clause too… this isn’t so bad with find/going thrpugh one file at a time, but it’s a nuisance for sure… we’re also trying to switch out all the creates for upserts at the same time and rewriting any logic we hand coded for batch queries into proper prisma batches like deleteMany, updateMany

we’re on grandfathered paid plans… support I think is still 24 hour email… im sure on the paid prisma plans it’s better, i think 9-5 email/phone is quoted in the higher plans…

also depends if you mean handholding or only for real issues… i agree handholding was better 1 year ago, but I mostly relied on the forums/github

re: prisma migrate… once that is all built and working.. then I still have to figure out how the clusterfuck of hosting works… 2 servers? 1 sql, 1 node server, plus also a prisma account? or what? it’s confusing to say the least… oh and then I need cloud functions for server side subscriptions to

we’ll probably do a write up afterwards

damn, yeah sounds tedious. thx for all the info @tsdexter. would love that writeup if you get to it

@tsdexter lol, it was a real issue for me but since it isn't in production, probably not a real issue in the grand scheme of things.

@tsdexter are you hosting anything yourself or serverless?

hey, are others seeing 504 errors on their endpoints?

i'm seeing this on my endpoints and from the graph.cool playground

@tsdexter i seem to be hitting the 504 as well

@justSlone My endpoints aren’t returning either

in my app i get TypeError: failed to fetch, via postman i get a cloudfront HTML doc that basically says "CloudFront attempted to establish a connection with the origin, but either the attempt failed or the origin closed the connection."

https://status.graph.cool seems to think everything is fine 😞

I am getting TypeError: Failed to fetch even in the console of graph.cool

yep me too 😞

There is definitely something wrong but status hasn’t been updated to reflect

My app won’t load for my users! 😞

yep me too

status.graph.cool is hardly accurate.. never has been.. not updated until a person at the office becomes aware, it’s not automated

graph.cool was sooo flawless for 6 months, but now i'm down with no way to fix it.

the berlin office should have people starting work in 2-3 hours

@chameleon im hosting a react/node app in google app engine that talks to graphcool… once prisma migrate is done, I will likely host the DB in google cloud sql, the app layer on app engine and the subscriptions with google cloud functions… I also am rebuilding the graphcool file storage api with google cloud storage and cloud functions

is it down for paying users too? I can only complain so much, i'm on free plan.

I didn't know about status.graph.cool The fact that it shows everything up crazy.

it is down accross the board, all my paid and free accounts in EU and US

even if status.graph.cool was automated, this is a tricky error to auto-detect. Gateway error could mean their servers aren't even seeing the requests to know they are failing. (tho an external network heartbeat would easily find this).