a dashboard & server for handling tracing results sent by a graphql endpoint
You have to create the listener service (with a dashboard) in a new project.
run npm i --save graphql-vision
in your listener project.
in your index.ts file:
import VisionServer from 'graphql-vision';
const visionServer = new VisionServer();
visionServer.run({port: 4000, dbOptions: {
type: "postgres",
host: "localhost",
port: 5432,
username: "postgres",
password: "postgres",
database: "apollo-tracing"
}});
- The
port
argument stands for the port that the vision server will be running on. - The
dbOptions
argument stands for the type-orm configurations. The vision server requires a DB for saving the tracing results.
It will create a listener server with a graphql endpoint and a dashboard.
You have to configure your graphql service to send the apollo tracing objects to the listener service.
You can use the simpler way, using the graphql-vision-plugin, or the manual way:
For example, we configured our graphql service (created by apollo server):
const trace = `mutation($tracing: TracerInput) {
addTracing(tracing: $tracing)
}`;
const server = new ApolloServer({
typeDefs, resolvers, tracing: true, plugins: [{
requestDidStart({}){
return {
willSendResponse({response}){
request('http://localhost:4000/graphql', trace, {tracing: response.extensions.tracing}).then(()=>{
console.log('success');
});
}
}
}
}]
});
The tracing:true
makes apollo server to return apollo tracing in response's extensions. Then we send these objects to the listener service created earlier, with url http://localhost:4000/graphql
.
- Clone the repository
npm i
cd src/dashboard && npm i
You can now run the example project with the following command:
npm run example