504 - Time outs

graphcool-cloud

#1

Seeing an increase in the number of 504 time out responses from a simple un-permissioned query against the Simple API endpoint. Started to see degraded performance as of a few days ago.


#2

Seeing this as well. Have reached out to support.


#4

Still happening frequently today.

Suck at a point where if paying would help, I would pay (we’re a big company about to migrate on, so that’s always been the plan), but I won’t pay because of the lack of acknowledgement is making me worried about the migration. Same thing with support. Says ‘answers usually in minutes’, still have unseen requests days later. I get that free support isn’t a huge priority but does that get better with a paid plan?


#5

@Graphcool liked my tweet about the issue and gave no additional details so :man_shrugging:

Hopefully that means they are working on it.


#6

Thanks for your feedback, @magus and @Aubron.
Depending on server load, queries may timeout and result in a 504 at times. We are currently improving stability and performance of Graphcool Cloud to alleviate this phenomenon.

You might be interested in getting a private Graphcool Framework cluster: https://www.graph.cool/cloud/
Also, you might be interested in Prisma: https://www.prismagraphql.com/

Here’s a comparison of Prisma and Graphcool Framework: Graphcool Framework and Prisma

Hope that helps!


#7

So, to be clear, there is no increased SLA on the paid plans for any tiers of Shared Clusters, only the private cluster and/or self deployment options, yes?

Sounds like Prisma Cloud is our ideal solution when that comes out, we’re migrating from Docker cloud anyway. Thanks for your help.