Frontend

The frontend is responsible to provide a UI for users, as well as an API leveraged by the UI or other clients. The application is based on Next.JS to provide a React web-based UI and API routes.

Frontend source

Server Instrumentation

It is recommended to use a Node required module when starting your NodeJS application to initialize the SDK and auto-instrumentation. When initializing the OpenTelemetry Node.js SDK, you optionally specify which auto-instrumentation libraries to leverage, or make use of the getNodeAutoInstrumentations() function which includes most popular frameworks. The utils/telemetry/Instrumentation.js file contains all code required to initialize the SDK and auto-instrumentation based on standard OpenTelemetry environment variables for OTLP export, resource attributes, and service name.

const opentelemetry = require('@opentelemetry/sdk-node');
const {
  getNodeAutoInstrumentations,
} = require('@opentelemetry/auto-instrumentations-node');
const {
  OTLPTraceExporter,
} = require('@opentelemetry/exporter-trace-otlp-grpc');
const {
  OTLPMetricExporter,
} = require('@opentelemetry/exporter-metrics-otlp-grpc');
const { PeriodicExportingMetricReader } = require('@opentelemetry/sdk-metrics');
const {
  alibabaCloudEcsDetector,
} = require('@opentelemetry/resource-detector-alibaba-cloud');
const {
  awsEc2Detector,
  awsEksDetector,
} = require('@opentelemetry/resource-detector-aws');
const {
  containerDetector,
} = require('@opentelemetry/resource-detector-container');
const { gcpDetector } = require('@opentelemetry/resource-detector-gcp');
const {
  envDetector,
  hostDetector,
  osDetector,
  processDetector,
} = require('@opentelemetry/resources');

const sdk = new opentelemetry.NodeSDK({
  traceExporter: new OTLPTraceExporter(),
  instrumentations: [
    getNodeAutoInstrumentations({
      // only instrument fs if it is part of another trace
      '@opentelemetry/instrumentation-fs': {
        requireParentSpan: true,
      },
    }),
  ],
  metricReader: new PeriodicExportingMetricReader({
    exporter: new OTLPMetricExporter(),
  }),
  resourceDetectors: [
    containerDetector,
    envDetector,
    hostDetector,
    osDetector,
    processDetector,
    alibabaCloudEcsDetector,
    awsEksDetector,
    awsEc2Detector,
    gcpDetector,
  ],
});

sdk.start();

Node required modules are loaded using the --require command line argument. This can be done in the scripts.start section of package.json and starting the application using npm start.

  "scripts": {
    "start": "node --require ./Instrumentation.js server.js",
  },

Traces

Span Exceptions and status

You can use the span object's recordException function to create a span event with the full stack trace of a handled error. When recording an exception also be sure to set the span's status accordingly. You can see this in the catch block of the NextApiHandler function in the utils/telemetry/InstrumentationMiddleware.ts file.

span.recordException(error as Exception);
span.setStatus({ code: SpanStatusCode.ERROR });

Create new spans

New spans can be created and started using Tracer.startSpan("spanName", options). Several options can be used to specify how the span can be created.

• root: true will create a new trace, setting this span as the root.
• links are used to specify links to other spans (even within another trace) that should be referenced.
• attributes are key/value pairs added to a span, typically used for application context.

span = tracer.startSpan(`HTTP ${method}`, {
  root: true,
  kind: SpanKind.SERVER,
  links: [{ context: syntheticSpan.spanContext() }],
  attributes: {
    'app.synthetic_request': true,
    [SemanticAttributes.HTTP_TARGET]: target,
    [SemanticAttributes.HTTP_STATUS_CODE]: response.statusCode,
    [SemanticAttributes.HTTP_METHOD]: method,
    [SemanticAttributes.HTTP_USER_AGENT]: headers['user-agent'] || '',
    [SemanticAttributes.HTTP_URL]: `${headers.host}${url}`,
    [SemanticAttributes.HTTP_FLAVOR]: httpVersion,
  },
});

Browser Instrumentation

The web-based UI that the frontend provides is also instrumented for web browsers. OpenTelemetry instrumentation is included as part of the Next.js App component in pages/_app.tsx. Here instrumentation is imported and initialized.

import FrontendTracer from '../utils/telemetry/FrontendTracer';

if (typeof window !== 'undefined') FrontendTracer();

The utils/telemetry/FrontendTracer.ts file contains code to initialize a TracerProvider, establish an OTLP export, register trace context propagators, and register web specific auto-instrumentation libraries. Since the browser will send data to an OpenTelemetry collector that will likely be on a separate domain, CORS headers are also setup accordingly.

As part of the changes to carry over the synthetic_request attribute flag for the backend services, the applyCustomAttributesOnSpan configuration function has been added to the instrumentation-fetch library custom span attributes logic that way every browser-side span will include it.

import {
  CompositePropagator,
  W3CBaggagePropagator,
  W3CTraceContextPropagator,
} from '@opentelemetry/core';
import { WebTracerProvider } from '@opentelemetry/sdk-trace-web';
import { SimpleSpanProcessor } from '@opentelemetry/sdk-trace-base';
import { registerInstrumentations } from '@opentelemetry/instrumentation';
import { getWebAutoInstrumentations } from '@opentelemetry/auto-instrumentations-web';
import { Resource } from '@opentelemetry/resources';
import { SemanticResourceAttributes } from '@opentelemetry/semantic-conventions';
import { OTLPTraceExporter } from '@opentelemetry/exporter-trace-otlp-http';

const FrontendTracer = async () => {
  const { ZoneContextManager } = await import('@opentelemetry/context-zone');

  const provider = new WebTracerProvider({
    resource: new Resource({
      [SemanticResourceAttributes.SERVICE_NAME]:
        process.env.NEXT_PUBLIC_OTEL_SERVICE_NAME,
    }),
  });

  provider.addSpanProcessor(new SimpleSpanProcessor(new OTLPTraceExporter()));

  const contextManager = new ZoneContextManager();

  provider.register({
    contextManager,
    propagator: new CompositePropagator({
      propagators: [
        new W3CBaggagePropagator(),
        new W3CTraceContextPropagator(),
      ],
    }),
  });

  registerInstrumentations({
    tracerProvider: provider,
    instrumentations: [
      getWebAutoInstrumentations({
        '@opentelemetry/instrumentation-fetch': {
          propagateTraceHeaderCorsUrls: /.*/,
          clearTimingResources: true,
          applyCustomAttributesOnSpan(span) {
            span.setAttribute('app.synthetic_request', 'false');
          },
        },
      }),
    ],
  });
};

export default FrontendTracer;

Metrics

TBD

Logs

TBD

Baggage

OpenTelemetry Baggage is leveraged in the frontend to check if the request is synthetic (from the load generator). Synthetic requests will force the creation of a new trace. The root span from the new trace will contain many of the same attributes as an HTTP request instrumented span.

To determine if a Baggage item is set, you can leverage the propagation API to parse the Baggage header, and leverage the baggage API to get or set entries.

    const baggage = propagation.getBaggage(context.active());
    if (baggage?.getEntry("synthetic_request")?.value == "true") {...}