Send Data with the OpenTelemetry Go SDK

Use the OpenTelemetry Go SDK to instrument Go applications in a standard, vendor-agnostic, and future-proof way and send telemetry data to Honeycomb.

In this guide, we will walk you through instrumenting with OpenTelemetry for Go, which will include adding automatic instrumentation to your application.

Before You Begin 

Before you can set up automatic instrumentation for your Go application, you will need to do a few things.

Prepare Your Development Environment 

To complete the required steps, you will need:

  • A working Go environment
  • An application written in Go

Get Your Honeycomb API Key 

To send data to Honeycomb, you’ll need to sign up for a free Honeycomb account and create a Honeycomb Ingest API Key. To get started, you can create a key that you expect to swap out when you deploy to production. Name it something helpful, perhaps noting that it’s a getting started key. Make note of your API key; for security reasons, you will not be able to see the key again, and you will need it later!

Tip
For setup, make sure you check the “Can create datasets” checkbox so that your data will show up in Honeycomb. Later, when you replace this key with a permanent one, you can uncheck that box.

If you want to use an API key you previously stored in a secure location, you can also look up details for Honeycomb API Keys any time in your Environment Settings, and use them to retrieve keys from your storage location.

Configure OpenTelemetry SDK 

To configure the OpenTelemetry SDK and enable automatic instrumentation of HTTP requests in your application, you will add the following packages to your application.

Acquire Dependencies 

Install OpenTelemetry Go packages:

go get \
  github.com/honeycombio/otel-config-go/otelconfig \
  go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp

Initialize 

Prepare your application to send spans to Honeycomb.

Open or create a file called main.go:

package main

import (
    "fmt"
    "log"
    "net/http"

    "github.com/honeycombio/otel-config-go/otelconfig"
    "go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp"
)

// Implement an HTTP Handler function to be instrumented
func httpHandler(w http.ResponseWriter, r *http.Request) {
    fmt.Fprintf(w, "Hello, World")
}

func main() {
    // use otelconfig to set up OpenTelemetry SDK
    otelShutdown, err := otelconfig.ConfigureOpenTelemetry()
    if err != nil {
        log.Fatalf("error setting up OTel SDK - %e", err)
    }
    defer otelShutdown()

    // Initialize HTTP handler instrumentation
    handler := http.HandlerFunc(httpHandler)
    wrappedHandler := otelhttp.NewHandler(handler, "hello")
    http.Handle("/hello", wrappedHandler)

    // Serve HTTP server
    log.Fatal(http.ListenAndServe(":3030", nil))
}

Configure Environment Variables 

Once you have acquired the necessary dependencies, you can configure your SDK to send events to Honeycomb, and then run your application to see traces.

export OTEL_SERVICE_NAME="your-service-name"
export OTEL_EXPORTER_OTLP_ENDPOINT="https://api.honeycomb.io:443" # US instance
#export OTEL_EXPORTER_OTLP_ENDPOINT="https://api.eu1.honeycomb.io:443" # EU instance
export OTEL_EXPORTER_OTLP_HEADERS="x-honeycomb-team=your-api-key"
Variable Description
OTEL_EXPORTER_OTLP_ENDPOINT Honeycomb endpoint to which you want to send your data.
OTEL_EXPORTER_OTLP_HEADERS Header containing x-honeycomb-team=, plus your API Key generated in Honeycomb.
OTEL_SERVICE_NAME Service name. When you send data, Honeycomb creates a dataset in which to store your data and uses this as the name. Can be any string.
Note

If you use Honeycomb Classic, you must specify the Dataset using the HONEYCOMB_DATASET environment variable.

export OTEL_EXPORTER_OTLP_HEADERS="x-honeycomb-team=your-api-key,x-honeycomb-dataset=your-dataset"

Run 

Run your application:

go run YOUR_APPLICATION_NAME.go

Be sure to replace YOUR_APPLICATION_NAME with the name of your application’s main file.

In Honeycomb’s UI, you should now see your application’s incoming requests and outgoing HTTP calls generate traces.

Add Custom Instrumentation 

Automatic instrumentation is the easiest way to get started with instrumenting your code. To get additional insight into your system, you should also add custom, or manual, instrumentation where appropriate. Follow the instructions below to add custom instrumentation to your code.

To learn more about custom, or manual, instrumentation, visit the comprehensive set of topics covered by Manual Instrumentation for Go in OpenTelemetry’s documentation.

Acquire a Tracer 

To create spans, you need to acquire a Tracer.

import (
  // ...
  "go.opentelemetry.io/otel"
  // ...
)

// ...
tracer := otel.Tracer("tracer.name.here")

When you create a Tracer, OpenTelemetry requires you to give it a name as a string. This string is the only required parameter.

When traces are sent to Honeycomb, the name of the Tracer is turned into the library.name field, which can be used to show all spans created from a particular tracer.

In general, pick a name that matches the appropriate scope for your traces. If you have one tracer for each service, then use the service name. If you have multiple tracers that live in different “layers” of your application, then use the name that corresponds to that “layer”.

The library.name field is also used with traces created from instrumentation libraries.

Add Attributes to Spans 

Adding context to a currently executing span in a trace can be useful. For example, you may have an application or service that handles users, and you want to associate the user with the span when querying your dataset in Honeycomb. To do this, get the current span from the context and set an attribute with the user ID. This example assumes you are writing a web application with the net/http package:

import (
  // ...
  "go.opentelemetry.io/otel/attribute"
  "go.opentelemetry.io/otel/trace"
  // ...
)

// ...
handler := func(w http.ResponseWriter, r *http.Request) {
  user := someServiceCall() // get the currently logged in user
  ctx := r.Context()
  span := trace.SpanFromContext(ctx)
  span.SetAttributes(attribute.Int("user.id", user.getID()))
}
// ...

This will add a user.id field to the current span, so you can use the field in WHERE, GROUP BY, or ORDER clauses in the Honeycomb query builder.

Create Spans 

To get the full picture of what is happening, you can leverage manual instrumentation to create custom spans that describe what is happening in your application. To do this, grab the tracer from the OpenTelemetry API:

import (
  // ...
  "go.opentelemetry.io/otel"
  // ...
)

// ...
tracer := otel.Tracer("my-app") // if not already in scope
ctx, span := tracer.Start(ctx, "expensive-operation")
defer span.End()
// ...

Add Multi-Span Attributes 

Sometimes you want to add the same attribute to many spans within the same trace. This attribute may include variables calculated during your program, or other useful values for correlation or debugging purposes.

To add this attribute to multiple spans, leverage the OpenTelemetry concept of baggage. Baggage allows you to add a key with a value as an attribute to every subsequent child span of the current application context, as long as you configured a BaggageSpanProcessor when you initialized OpenTelemetry.

First, install the baggagetrace package in your terminal:

go get go.opentelemetry.io/contrib/processors/baggage/baggagetrace

Then, when configuring the OpenTelemetry SDK tracer provider, add the baggage span processor:

import (
    // ...
    "go.opentelemetry.io/contrib/processors/baggage/baggagetrace"
    // ...
)

  // Create a new tracer provider with the baggage span processor
  tp := trace.NewTracerProvider(
    baggagetrace.New()
    // ...
  )

Finally, add a baggage entry for the current trace and replace key and value with your desired key-value pair:

import (
  // ...
  "go.opentelemetry.io/otel/attribute"
  "go.opentelemetry.io/otel/trace"
  "go.opentelemetry.io/otel/baggage"
  // ...
)

// ...
handler := func(w http.ResponseWriter, r *http.Request) {
  ctx := r.Context()

  // add the user ID attribute to baggage and create new context
  bag := baggage.FromContext(ctx)
  multiSpanAttribute, _ := baggage.NewMember("key", "value")
  bag, _ = bag.SetMember(multiSpanAttribute)
  ctx = baggage.ContextWithBaggage(ctx, bag)

  tracer := otel.Tracer("my-app") // if not already in scope

  // every subsequent span created from this context, and any of its child spans,
  // will have the user ID attribute from baggage
  ctx, span := tracer.Start(ctx, "expensive-operation")
  defer span.End()
}
// ...

Note: Any Baggage attributes that you set in your application will be attached to outgoing network requests as a header. If your service communicates to a third party API, do NOT put sensitive information in the Baggage attributes.

Automatic Instrumentation using eBPF 

To instrument http and gRPC requests in Go, usually you must wrap requests with OpenTelemetry instrumentation libraries. However, a new project allows for automatic instrumentation of http and gRPC requests using eBPF, which requires no application code changes.

Because the automatic instrumentation uses eBPF, it requires a Linux kernel. Automatic instrumentation should work on any Linux kernel above 4.4. For most cloud-native applications, this means you must include the Docker image to run as an agent in a container for each application.

Configure 

You must configure the following options for each instrumented application:

  • Your application path, specified by OTEL_GO_AUTO_TARGET_EXE, is where the agent will watch for processes.
  • Your endpoint, specified by OTEL_EXPORTER_OTLP_ENDPOINT, is where telemetry will be sent.
  • Your service name, specified by OTEL_SERVICE_NAME, will be used as the Service Dataset in Honeycomb, which is where data is stored.

Configure an OpenTelemetry Collector to receive traces over OTLP/gRPC and export those traces to Honeycomb. Then follow the instructions for deployment in Kubernetes or from source on a Linux machine.

Running in Kubernetes 

The automatic instrumentation agent runs in the same container node as your application. The agent requires shareProcessNamespace, as well as some elevated permissions in securityContext.

The following example shows what a deployment spec.template.spec could look like with an existing application called “my-service”:

    spec:
      shareProcessNamespace: true
      securityContext: {}
      terminationGracePeriodSeconds: 30
      containers:
        - name: my-service
          image: my-service:v42
          ports:
          - containerPort: 7007
            name: http
        - name: my-service-instrumentation
          image: ghcr.io/open-telemetry/opentelemetry-go-instrumentation/autoinstrumentation-go:v0.2.0-alpha
          env:
          - name: OTEL_GO_AUTO_TARGET_EXE
            value: /app/my-service
          - name: OTEL_EXPORTER_OTLP_ENDPOINT
            value: http://otel-collector:4317
          - name: OTEL_SERVICE_NAME
            value: my-service-name
          securityContext:
            runAsUser: 0
            capabilities:
              add:
                - SYS_PTRACE
            privileged: true

Example 

If you prefer to learn by example, we provide an example application that illustrates a Kubernetes deployment.

Running on a Linux Machine 

If the application is running in Linux, an alternative to Kubernetes is to build and run the instrumentation from source. To use the instrumentation without a Docker image, build a binary from source and save as otel-go-instrumentation.

Set environment variables for the application, service name, and endpoint, and pass into a run command with the instrumentation.

The following example shows how to enable instrumentation for an application running in ~/app/my-service:

OTEL_GO_AUTO_TARGET_EXE=~/app/my-service \ # application being instrumented
  OTEL_SERVICE_NAME=my-service \ # name of service in telemetry data
  OTEL_EXPORTER_OTLP_ENDPOINT=http://otel-collector:4317 \ # send to collector
  ./otel-go-instrumentation
Note
Because eBPF has powerful capabilities, running this instrumentation may require additional privileges on the host, such as running with the sudo command.

Sampling 

You can configure the OpenTelemetry SDK to sample the data it generates. Honeycomb weights sampled data based on sample rate, so you must set a resource attribute containing the sample rate.

Use a TraceIdRatioBased sampler, with a ratio expressed as 1/N. Then, also create a resource attribute called SampleRate with the value of N. This allows Honeycomb to reweigh scalar values, like counts, so that they are accurate even with sampled data.

In the example below, our goal is to keep approximately half (1/2) of the data volume. The resource attribute contains the denominator (2), while the OpenTelemetry sampler argument contains the decimal value (0.5).

export OTEL_TRACES_SAMPLER="traceidratio"
export OTEL_TRACES_SAMPLER_ARG=0.5
export OTEL_RESOURCE_ATTRIBUTES="SampleRate=2"

Choosing between gRPC and HTTP 

Most OpenTelemetry SDKs have an option to export telemetry as OTLP either over gRPC or HTTP/protobuf, with some also offering HTTP/JSON. If you are trying to choose between gRPC and HTTP, keep in mind:

  • Some SDKs default to using gRPC, and it may be easiest to start with the default option.
  • Some firewall policies are not set up to handle gRPC and require using HTTP.
  • gRPC may improve performance, but its long-lived connections may cause problems with load balancing, especially when using Refinery.

gRPC default export uses port 4317, whereas HTTP default export uses port 4318.

Endpoint URLs for OTLP/HTTP 

When using the OTEL_EXPORTER_OTLP_ENDPOINT environment variable with an SDK and an HTTP exporter, the final path of the endpoint is modified by the SDK to represent the specific signal being sent.

For example, when exporting trace data, the endpoint is updated to append v1/traces. When exporting metrics data, the endpoint is updated to append v1/metrics. So, if you were to set the OTEL_EXPORTER_OTLP_ENDPOINT to https://api.honeycomb.io, traces would be sent to https://api.honeycomb.io/v1/traces and metrics would be sent to https://api.honeycomb.io/v1/metrics.

The same modification is not necessary for gRPC.

export OTEL_EXPORTER_OTLP_ENDPOINT=https://api.honeycomb.io # US instance
#export OTEL_EXPORTER_OTLP_ENDPOINT=https://api.eu1.honeycomb.io # EU instance

If the desired outcome is to send data to a different endpoint depending on the signal, use OTEL_EXPORTER_OTLP_<SIGNAL>_ENDPOINT instead of the more generic OTEL_EXPORTER_OTLP_ENDPOINT.

When using a signal-specific environment variable, these paths must be appended manually. Set OTEL_EXPORTER_OTLP_TRACES_ENDPOINT for traces, appending the endpoint with v1/traces, and OTEL_EXPORTER_OTLP_METRICS_ENDPOINT for metrics, appending the endpoint with v1/metrics.

Send both traces and metrics to Honeycomb using this method by setting the following variables:

export OTEL_EXPORTER_OTLP_TRACES_ENDPOINT=https://api.honeycomb.io/v1/traces # US instance
#export OTEL_EXPORTER_OTLP_TRACES_ENDPOINT=https://api.eu1.honeycomb.io/v1/traces # EU instance

export OTEL_EXPORTER_OTLP_METRICS_ENDPOINT=https://api.honeycomb.io/v1/metrics # US instance
#export OTEL_EXPORTER_OTLP_METRICS_ENDPOINT=https://api.eu1.honeycomb.io/v1/metrics # EU instance

More details about endpoints and signals can be found in the OpenTelemetry Specification.

Troubleshooting 

To explore common issues when sending data, visit Common Issues with Sending Data in Honeycomb.