I have a conceptual question that someone with experience running graphcool in production or someone working for graphcool will hopefully know.
I have a Recipe type which has a many to many relationship with the Ingredient type and want to implement querying recipes by their ingredients so I can compute a list of recipes that will be easy to make with a given list of ingredients. The schema is basically already set up as an inverted index so the concept is sound but I wonder how it will scale.
Are there any known upper thresholds to payload size? Say I end up querying a list of 30 ingredients that each have thousands of recipes connected. Or will the shared and private cluster memory be the only limiting factor.
I know I can turn to things like Elasticsearch to linearly scale this type of application but I would like to know how far I could push the graphcool service.