Validation rules can be used to validate queries before they are passed to the execution engine and by this save execution time. Validation results are cached with the query so that validation rules are only run once per query. In query validation rules you can access the query and the schema but the argument values and the variable values are not yet coerced. If you need access to those a field middleware might be better suited.

IQueryValidationRule

The rule interface itself is simple, basically the validation middleware will call validate and pass in the schema and the parsed query.

public interface IQueryValidationRule
{
    QueryValidationResult Validate(ISchema schema, DocumentNode query);
}

QuerySyntaxWalker

If your validation is just a simple lookup, then you could just try to do that against the DocumentNode for everything where you would have to traverse the query graph or where you want to access only certain kinds of syntax nodes we recommend using the QuerySyntaxWalker<T>.

Important: In your validation rules only validate your own errors and ignore other errors.

Let us say we want to ensure a certain rule for every field selection in a query, then we could write the following QuerySyntaxWalker<T>.

internal sealed class MaxDepthVisitor
    : QuerySyntaxWalker<object>
{
    protected override void VisitField(
        FieldNode field,
        object context)
    {
        // field validation code

        base.VisitField(field, context
    }
}

The type parameter defines the visitor context type. The context type can be used to pass a visitation context between the visit methods.

Add Validation Rules

Validation rules can be added via the QueryExecutionBuilder like the following:

QueryExecutionBuilder
    .New()
    .AddValidationRule<MyCustomRule>()
    .UseDefaultPipeline()
    .Build(schema);

Since the validation rules are instantiated like a query middleware, you can only access services defined by the QueryExecutionBuilder.

Blogs

Guarding against N+1 issues in GraphQL