otlp

OTLP Exporter Example

This is an example of how to use the OpenTelemetry Protocol (OTLP) exporter.

Traces

The application in grpc_main.cc initializes an OtlpGrpcExporter instance, the application in http_main.cc initializes an OtlpHttpExporter instance. And they register a tracer provider from the OpenTelemetry SDK. The application then calls a foo_library which has been instrumented using the OpenTelemetry API.

Logs

The application in http_log_main.cc initializes an OtlpHttpLogRecordExporter instance, the application in grpc_log_main.cc initializes an OtlpGrpcLogRecordExporter instance. And they register a logger provider from the OpenTelemetry SDK. The application then calls a logs_foo_library which has been instrumented using the OpenTelemetry API.

Metrics

The application in grpc_metrics_main.cc initializes an OtlpGrpcMetricExporter instance. And it registers a meter provider from the OpenTelemetry SDK. The application then calls a metrics_foo_library which has been instrumented using the OpenTelemetry API.

Enable SSL/TLS

To enable TLS authentication for OTLP grpc exporter, SslCredentials can be used by specifying the path to client certificate pem file, or the string containing this certificate via OtlpGrpcExporterOptions. The path to such a .pem file can be provided as a command-line argument along with the collector endpoint to the main binary invocation above.

Running OpenTelemetry Collector as docker container

Resulting spans are exported to the OpenTelemetry Collector using the OTLP exporter. The OpenTelemetry Collector can be configured to export to other backends (see list of supported backends).

Follow below command to run the OpenTelemetry Collector with OTLP receiver in docker which dumps the received data into console. See Getting Started for more information.

Open a terminal window at the root directory of this repo and launch the OpenTelemetry Collector with an OTLP receiver by running:

• On Unix based systems use:

docker run --rm -it -p 4317:4317 -p 4318:4318 -v $(pwd)/examples/otlp:/cfg otel/opentelemetry-collector:0.59.0 --config=/cfg/opentelemetry-collector-config/config.dev.yaml

• On Windows use:

docker run --rm -it -p 4317:4317 -p 4318:4318 -v "%cd%/examples/otlp":/cfg otel/opentelemetry-collector:0.59.0 --config=/cfg/opentelemetry-collector-config/config.dev.yaml

Note that the OTLP gRPC and HTTP exporters connects to the Collector at localhost:4317 and localhost:4318/v1/traces respectively. This can be changed with first argument from command-line, for example: ./example_otlp_grpc gateway.docker.internal:4317 and ./example_otlp_http gateway.docker.internal:4318/v1/traces.

Once you have the Collector running, see CONTRIBUTING.md for instructions on building and running the example.

Additional notes regarding Abseil library

gRPC internally uses a different version of Abseil than OpenTelemetry C++ SDK.

One option to optimize your code is to build the SDK with system-provided Abseil library. If you are using CMake, then -DWITH_ABSEIL=ON may be passed during the build of SDK to reuse the same Abseil library as gRPC. If you are using Bazel, then --@io_opentelemetry_cpp/api:with_abseil=true may be passed to reuse your Abseil library in your project.

If you do not want to pursue the above option, and in case if you run into conflict between Abseil library and OpenTelemetry C++ absl::variant implementation, please include either grpcpp/grpcpp.h or opentelemetry/exporters/otlp/otlp_grpc_exporter.h BEFORE any other API headers. This approach efficiently avoids the conflict between the two different versions of Abseil.