Skip to content

Latest commit

 

History

History
299 lines (238 loc) · 7.41 KB

README.md

File metadata and controls

299 lines (238 loc) · 7.41 KB

Mercurius Explain

A Mercurius plugin that exports some execution info in a query.

The additional information currently exported by the plugin are:

  • profiling of resolvers execution time
  • number of call per resolver

The information is added to the extensions.explain attribute in the GQL response.

{
  extensions: {
    explain: {
      version: '1.1.1', // The version in package.json
      profiler: {
        ...
      },
      resolverCalls: {
        ...
      }
    }
  }
}

Profiler

The profiler contains the execution time of each resolver called.

  • data is an array with the definition of the profiler entry:
    • path is a string that represents the subpath of the resolver
    • begin is number that represents the start time in NANOSECONDS
    • end is number that represents the end time in NANOSECONDS
    • time is number that represents the time between begin and end in NANOSECONDS

Example:

{
  extensions: {
    explain: {
      profiler: {
        data: [
          {
            path: 'user',
            begin: 100, // time in nanoseconds,
            end: 300, // time in nanoseconds,
            time: 200, // time in nanoseconds,
          },
          {
            path: 'user.address',
            begin: 301, // time in nanoseconds,
            end: 400, // time in nanoseconds,
            time: 99, // time in nanoseconds,
          },
          {
            path: 'user.status',
            begin: 301, // time in nanoseconds,
            end: 350, // time in nanoseconds,
            time: 49, // time in nanoseconds,
          },
          ...
        ]
      }
      ...
    }
  }
}

Resolver Calls

Every time a resolver is invoked, a counter keeps track of the call and returns a report with resolverCalls object:

  • data is an array that contains an object for each resolver that has been called:
    • key is a string that define the resolver.
    • count is a number that define the number of calls for a resolver.

example

{
  extensions: {
    explain: {
      resolverCalls: {
        data:data: [
          { key: "Query.users", count: 1 },
          { key: "User.status", count: 2 },
          { key: "User.contacts", count: 2 },
          { key: "Contact.emails", count: 2 }
        ]
      },
      ...
    }
  }
}

Install

npm i fastify mercurius mercurius-explain

Quickstart

import Fastify from 'fastify'
import mercurius from 'mercurius'
import explain, { explainGraphiQLPlugin } from 'mercurius-explain'

const app = Fastify({ logger: true })

const schema = `
  type Query {
    add(x: Int, y: Int): Int
  }
`

const resolvers = {
  Query: {
    async add(_, { x, y }, { reply }) {
      return x + y
    }
  }
}

app.register(mercurius, {
  schema,
  resolvers,
  graphiql: {
    enabled: true,
    plugins: [explainGraphiQLPlugin()]
  }
})

app.register(explain, {})

app.listen({ port: 3000 })

Test:

curl -X POST -H 'content-type: application/json' -d '{ "query": "{ add(x: 2, y: 2) }" }' localhost:3000/graphql

Options

  • enabled: boolean | function (schema, source, context) => boolean. Enables or disables the data collection and the enrichment of the response. By default the action is enabled.

If federated: true, this field will be ignored and be enabled by default.

Examples:

// Data enrichment disabled
app.register(explain, {
   enabled: false
}
// Data are collected and returned only if the request has 'x-mercurius-explain' header
app.register(explain, {
  enabled: ({ schema, source, context }) =>
    context.reply.request.headers['x-mercurius-explain']
})
  • gateway: boolean. If the application is a Mercurius Gateway enables or disables the collection of extensions.explain from the services that are part of the gateway that use mercurius-explain. Use the helper function getExplainFederatedHeader in service.rewriteHead in the service definition to enabled federation mode like in the example.

    It is required to enable extensions collector in the gateway definition. If the federated services are in federation mode it is required to use rewriteHeaders to forward the getExplainFederatedHeader

Examples:

const gateway = Fastify()

await gateway.register(mercuriusGateway, {
  gateway: {
    services: [
      {
        name: 'foo',
        url: `http://localhost:3000/graphql`,
        collectors: {
          collectExtensions: true // enabled extensions collector required
        },
        rewriteHeaders: (headers, context) => {
          return { ...getExplainFederatedHeader(context) } // this will forward the federation header to service
        }
      },
      {
        name: 'bar',
        url: `http://localhost:3001/graphql`,
        collectors: {
          collectExtensions: true // enabled extensions collector required
        },
        rewriteHeaders: (headers, context) => {
          return { ...getExplainFederatedHeader(context) } // this will forward the federation header to service
        }
      }
    ]
  },
  ...opts
})

gateway.register(mercuriusExplain, {
  enabled: true,
  gateway: true // gateway enabled
})
  • federated: boolean. If the application is a federated service enables or disables federation mode. Federation mode allows the service to trust any request from the gateway. The application will check the presence of a specific header in the request from the gateway and will ignore the enabled field.

Example:

  • Federated service
fastify.register(mercuriusExplain, {
  enabled: explainEnabled,
  federated: true // enabled federation
})

getExplainFederatedHeader helper

This function returns the headers required for federation mode.

getExplainFederatedHeader: function(context)

await gateway.register(mercuriusGateway, {
  gateway: {
    services: [
      {
        name: 'foo',
        url: `http://localhost:3000/graphql`,
        rewriteHeaders: (headers, context) => {
          return { ...getExplainFederatedHeader(context) } // this will forward the header to federated service
        }
      }
    ]
  }
})

Add the viewer plugin to mercurius GraphiQL (mercurius-explain-graphiql-plugin)

In mercurius it is possibile to add to the self hosted GraphiQL app the plugin mercurius-explain-graphiql-plugin to show the data returned by mercurius explain.

explainGraphiQLPlugin helper

This function return the required structure to initialize the plugin.

explainGraphiQLPlugin: function(options)

  • options: null | object
    • options.version: string. The version of the GraphiQL plugin to be loaded. Default: the same major version of the backend plugin

Example:

import { explainGraphiQLPlugin } from 'mercurius-explain'

app.register(mercurius, {
  schema,
  resolvers,
  graphiql: {
    enabled: true,
    plugins: [explainGraphiQLPlugin()]
  }
})

The explainGraphiQLPlugin function initializes by default the plugin with the same major version in the package.json (eg. if the package is 3.4.5 it will load the version ^3 of the GraphiQL plugin).

It's possible to override the version by passing a parameter.

...
plugins: [explainGraphiQLPlugin({version: '3.4.5')]

// or

plugins: [explainGraphiQLPlugin({version: '^4')]
...