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 can set up automatic instrumentation for your Go application, you will need to do a few things.
To complete the required steps, you will need:
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!
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.
To configure the OpenTelemetry SDK and enable automatic instrumentation of HTTP requests in your application, you will add the following packages to your application.
Install OpenTelemetry Go packages:
go get \
github.com/honeycombio/otel-config-go/otelconfig \
go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp
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))
}
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. |
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 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.
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.
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.
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.
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()
// ...
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.
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.
You must configure the following options for each instrumented application:
OTEL_GO_AUTO_TARGET_EXE, is where the agent will watch for processes.OTEL_EXPORTER_OTLP_ENDPOINT, is where telemetry will be sent.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.
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
If you prefer to learn by example, we provide an example application that illustrates a Kubernetes deployment.
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
sudo command. 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"
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:
gRPC default export uses port 4317, whereas HTTP default export uses port 4318.
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.
To explore common issues when sending data, visit Common Issues with Sending Data in Honeycomb.