Query Cost Analysis

[Sherlouk]

For applications which seek the ability to charge for access to their GraphQL API, a cost analysis must be done to charge request based on their impact. As in, requesting a single model will cost less than multiple models with relationships and such.

It would be good if we could provide a way of intercepting requests before they're executed. This would allow us to throw an error (also helpful for auth), record usage, or even add support for a "dryRun" mode which simply returns the query cost without actually executing it.

Examples:

https://github.com/pa-bru/graphql-cost-analysis

https://github.com/slicknode/graphql-query-complexity

[d-exclaimation]

It seems like the capability needed to be implemented first is validation rules. I will see what I can do.

[d-exclaimation]

The validation rules is already in the GraphQLSwift/GraphQL package which should allow something like https://github.com/pa-bru/graphql-cost-analysis.

However, the capability to provide custom validation rules are not public outside the GraphQLSwift/GraphQL package. I am not sure if that was an intentional decision by the maintainer for GraphQLSwift/GraphQL. I have asked about this and once that feature is made public, I will immediately update this package as well.

GraphQLSwift/GraphQL#98

[d-exclaimation]

In the mean time, you can technically use the context builder to intercept request by throwing an error before it is executed by the resolver which was updated in v0.6.0.

[Sherlouk]

I'll be sure to take a look at that and see if it works for my use case. Thanks 😊

[d-exclaimation]

Sorry, this feature took a while, but validation rules are now in Pioneer since v0.9.0. You should be able to provide custom rules that intercept operation before any resolver is executed.

You should also be capable of computing the validation rules for a specific GraphQLRequest (this is used by Query Cost Analysis to grab the variables which not provided in the ValidationContext).

The only thing missing now is an implementation of Query Cost Analysis or Query Complexity.