Hola. Yep banking on using Graphcool in a current client project. Will see me around alot 😎 . So we can only run 30 mutations per second with the shared cluster thus nested migrations can cause issues?

basically yea, but also if you run too many queries too quickly for example

I really think the mutations/writes should be put into a pipeline so a bigger nested mutation just takes a little longer but falls within the 30 mutations per second cap.

there is an internal queue for DB actions but even that might 'overheat' 🙂

@nilan whats the maximum mutations/writes possible on a private cluster

We do it in a local react app or server based react const .. and then pass the info / mutation through the router on a HoC .. reducing the nested calls to the api