Overview

There are a few reasons to manually instrument your applications with the OpenTelemetry API:

• You are not using Datadog supported library instrumentation.
• You want to extend the ddtrace library’s functionality.
• You need finer control over instrumenting your applications.

The ddtrace library provides several techniques to help you achieve these goals. The following sections demonstrate how to use the OpenTelemetry API for custom instrumentation to use with Datadog.

Setup

To configure OpenTelemetry to use the Datadog trace provider:

1. If you have not yet read the instructions for auto-instrumentation and setup, start with the Java Setup Instructions.

2. Make sure you only depend on the OpenTelemetry API (and not the OpenTelemetry SDK).

3. Set the dd.trace.otel.enabled system property or the DD_TRACE_OTEL_ENABLED environment variable to true.

Adding span tags

Add custom span tags

Add custom tags to your spans corresponding to any dynamic value within your application code such as customer.id.

import io.opentelemetry.api.trace.Span;

public void doSomething() {
  Span span = Span.current();
  span.setAttribute("user-name", "Some User");
}

Adding tags globally to all spans

The dd.tags property allows you to set tags across all generated spans for an application. This is useful for grouping stats for your applications, data centers, or any other tags you would like to see in Datadog.

java -javaagent:<DD-JAVA-AGENT-PATH>.jar \
    -Ddd.tags=datacenter:njc,<TAG_KEY>:<TAG_VALUE> \
    -jar <YOUR_APPLICATION_PATH>.jar

Setting errors on a root span from a child span

To set an error on a root span from a child span, you can use the setStatus method on the current span like this:

import static io.opentelemetry.api.trace.StatusCode.ERROR;
import io.opentelemetry.api.trace.Span;

public void doSomething() {
  Span span = Span.current();
  span.setStatus(ERROR, "Some error details...");
}

Setting tags and errors on a root span from a child span

This example demonstrates how to set tags and errors on a root span from a child span:

import io.opentelemetry.api.OpenTelemetry;
import io.opentelemetry.api.trace.Span;
import io.opentelemetry.api.trace.Tracer;
import io.opentelemetry.context.Context;
import io.opentelemetry.context.ContextKey;
import io.opentelemetry.context.Scope;
import io.opentelemetry.exporter.otlp.trace.OtlpGrpcSpanExporter;
import io.opentelemetry.sdk.OpenTelemetrySdk;
import io.opentelemetry.sdk.resources.Resource;
import io.opentelemetry.sdk.trace.SdkTracerProvider;
import io.opentelemetry.sdk.trace.export.BatchSpanProcessor;
import io.opentelemetry.semconv.ResourceAttributes;
import java.util.concurrent.TimeUnit;

public class Example {

  private final static ContextKey<Span> CONTEXT_KEY =
    ContextKey.named("opentelemetry-traces-local-root-span");

  public void begin() {
    tracer = GlobalOpenTelemetry.getTracer("my-scope", "0.1.0");
    Span parentSpan = tracer.spanBuilder("begin").startSpan();
    try (Scope scope = parentSpan.makeCurrent()) {
      createChildSpan();
    } finally {
      parentSpan.end();
    }
  }

  private void createChildSpan() {
    Span childSpan = tracer.spanBuilder("child-span").startSpan();
    try {
      Span rootSpan = Context.current().get(CONTEXT_KEY);
        if (null != rootSpan) {
          rootSpan.setAttribute("my-attribute", "my-attribute-value");
          rootSpan.setStatus(StatusCode.ERROR, "Some error details...");
        } 
    } finally {
      childSpan.end();
    }
  }

}

Adding spans

If you aren’t using a supported framework instrumentation, or you would like additional depth in your application’s traces, you may want to add custom instrumentation to your code for complete flame graphs or to measure execution times for pieces of code.

If modifying application code is not possible, use the environment variable dd.trace.methods to detail these methods.

If you have existing @Trace or similar annotations, or prefer to use annotations to complete any incomplete traces within Datadog, use Trace Annotations.

Traces may also be created using the OpenTelemetry @WithSpan annotation as described in Trace annotations.

Trace annotations

Add @WithSpan to methods to have them be traced when running OpenTelemetry and the dd-java-agent.jar. If the Agent is not attached, this annotation has no effect on your application. OpenTelemetry’s @WithSpan annotation is provided by the opentelemetry-instrumentation-annotations dependency.

import io.opentelemetry.instrumentation.annotations.WithSpan;

public class SessionManager {

  @WithSpan
  public static void saveSession() {
    // your method implementation here
  }
}

Manually creating a new span

To manually create new spans within the current trace context:

import io.opentelemetry.api.OpenTelemetry;
import io.opentelemetry.api.trace.Span;
import io.opentelemetry.api.trace.Tracer;
import io.opentelemetry.context.Scope;
import io.opentelemetry.exporter.otlp.trace.OtlpGrpcSpanExporter;
import io.opentelemetry.sdk.OpenTelemetrySdk;
import io.opentelemetry.sdk.resources.Resource;
import io.opentelemetry.sdk.trace.SdkTracerProvider;
import io.opentelemetry.sdk.trace.export.BatchSpanProcessor;
import io.opentelemetry.semconv.ResourceAttributes;
import java.util.concurrent.TimeUnit;

public class Example {

  public void doSomething() {
    Tracer tracer = GlobalOpenTelemetry.getTracer("my-scope", "0.1.0");
    Span span = tracer.spanBuilder("my-resource").startSpan();
    try (Scope scope = span.makeCurrent()) {
      // do some work
    } catch (Throwable t) {
      span.recordException(t);
      throw t;
    } finally {
      span.end();
    }
  }

}

Adding span events

You can add span events using the addEvent API. This method requires a name parameter and optionally accepts attributes and timestamp parameters. The method creates a new span event with the specified properties and associates it with the corresponding span.

• Name [required]: A string representing the event’s name.
• Attributes [optional]: Zero or more key-value pairs with the following properties:
  • The key must be a non-empty string.
  • The value can be either:
    • A primitive type: string, Boolean, or number.
    • A homogeneous array of primitive type values (for example, an array of strings).
  • Nested arrays and arrays containing elements of different data types are not allowed.
• Timestamp [optional]: A UNIX timestamp representing the event’s occurrence time. Expects an Instant object.

The following examples demonstrate different ways to add events to a span:

Attributes eventAttributes = Attributes.builder()
    .put(AttributeKey.longKey("int_val"), 1L)
    .put(AttributeKey.stringKey("string_val"), "two")
    .put(AttributeKey.longArrayKey("int_array"), Arrays.asList(3L, 4L))
    .put(AttributeKey.stringArrayKey("string_array"), Arrays.asList("5", "6"))
    .put(AttributeKey.booleanArrayKey("bool_array"), Arrays.asList(true, false))
    .build();

span.addEvent("Event With No Attributes");
span.addEvent("Event With Some Attributes", eventAttributes);

Read the OpenTelemetry specification for more information.

Recording exceptions

To record exceptions, use the recordException API. This method requires an exception parameter and optionally accepts a UNIX timestamp parameter. It creates a new span event that includes standardized exception attributes and associates it with the corresponding span.

The following examples demonstrate different ways to record exceptions:

span.recordException(new Exception("Error Message"));
span.recordException(new Exception("Error Message"), 
    Attributes.builder().put(AttributeKey.stringKey("status"), "failed").build());

Read the OpenTelemetry specification for more information.

Trace client and Agent configuration

Both the tracing client and Datadog Agent offer additional configuration options for context propagation. You can also exclude specific resources from sending traces to Datadog if you don’t want those traces to be included in calculated metrics, such as traces related to health checks.

Propagating context with headers extraction and injection

You can configure the propagation of context for distributed traces by injecting and extracting headers. Read Trace Context Propagation for information.

Resource filtering

Traces can be excluded based on their resource name, to remove synthetic traffic such as health checks from reporting traces to Datadog. This and other security and fine-tuning configurations can be found on the Security page or in Ignoring Unwanted Resources.

Further Reading

Documentation, liens et articles supplémentaires utiles:

Explore your services, resources, and tracesDOCUMENTATION Interoperability of OpenTelemetry API and Datadog instrumented tracesDOCUMENTATION