# Request Lifecycle
To understand how queries are executed, we can examine the lifecycle of an incoming GraphQL request. This section should provide an understanding of the involved steps, their order and purpose.
All requests to the configured GraphQL endpoint - usually
/graphql - are routed
to the single
# HTTP Middleware
We are still dealing with HTTP requests, so Laravel middleware applies as usual.
# Request Parsing
The incoming request is parsed into the elements that make a GraphQL request. Lighthouse aims to handle requests following the informal specification at graphql.org (opens new window).
# Schema Construction
The defined types are collected from the
.graphql schema files and programmatic definitions.
Schema transformations are applied, such as expansion of
Lighthouse assembles a runtime representation - the executable schema. This step is typically cached for larger schemas, to enhance performance.
# Query Validation
The GraphQL query is validated to ensure it matches the schema. Lighthouse makes sure the requested fields are available in the schema, and the correct variables are passed.
# Field Execution
Starting from the root level, the fields within the query are executed.
Each field may be wrapped with field middleware, which can add authentication, authorization, fine-grained validation, and more. Finally, the field resolver is called to produce a value for the field.
If the field contains a subselection, the same process happens for the nested fields, until we either abort with an error or traversed the entire query tree.
# Error Handling
If an error occurs during field execution, the traversal of that particular subtree is stopped. Lighthouse collects the errors to allow the rest of the query to execute.
# Result Assembly
The execution results are assembled into a structure that resembles the query of the client. Errors are properly formatted and included in the response. The response is sent to the client.