11 KiB
Building opentelemetry-cpp
CMake and Bazel are the official build systems for opentelemetry-cpp.
Dependencies
You can link OpenTelemetry C++ SDK with libraries provided in dependencies.md (complete list of libraries with versions used in our CI can be found here).
Build instructions using CMake
Prerequisites for CMake
- A supported platform (e.g. Windows, macOS or Linux). Refer to Platforms Supported for more information.
- A compatible C++ compiler supporting at least C++11. Major compilers are supported. Refer to Supported Compilers for more information.
- Git for fetching opentelemetry-cpp source code from repository. To install Git, consult the Set up Git guide on GitHub.
- CMake for building opentelemetry-cpp API, SDK with their unittests. We use CMake version 3.15.2 in our build system. To install CMake, consult the Installing CMake guide.
- GoogleTest framework to build and run the unittests. Refer to third_party_release for version of GoogleTest used in CI. To install GoogleTest, consult the GoogleTest Build Instructions.
- Google Benchmark framework to build and run benchmark tests. Refer to third_party_release for version of Benchmark used in CI. To install Benchmark, consult the GoogleBenchmark Build Instructions.
- Apart from above core requirements, the Exporters and Propagators have their build dependencies which are not covered here. E.g, Otlp Exporter needs grpc/protobuf library, Zipkin exporter needs nlohmann-json and libcurl, ETW exporter need nlohmann-json to build. This is covered in the build instructions for each of these components.
Building as standalone CMake Project
-
Getting the opentelementry-cpp source:
# Change to the directory where you want to create the code repository $ cd ~ $ mkdir source && cd source $ git clone --recursive https://github.com/open-telemetry/opentelemetry-cpp Cloning into 'opentelemetry-cpp'... ... Resolving deltas: 100% (3225/3225), done. $ -
Navigate to the repository cloned above, and create the
CMakebuild configuration.$ cd opentelemetry-cpp $ mkdir build && cd build $ cmake .. -- The C compiler identification is GNU 9.3.0 -- The CXX compiler identification is GNU 9.3.0 ... -- Configuring done -- Generating done -- Build files have been written to: /home/<user>/source/opentelemetry-cpp/build $Some of the available cmake build variables we can use during cmake configuration:
-DCMAKE_POSITION_INDEPENDENT_CODE=ON: Please note that with default configuration, the code is compiled without-fpicoption, so it is not suitable for inclusion in shared libraries. To enable the code for inclusion in shared libraries, this variable is used.-DBUILD_SHARED_LIBS=ON: To build shared libraries for the targets. Please refer to note below for Windows DLL support-DWITH_OTLP=ON: To enable building Otlp exporter.-DWITH_PROMETHEUS=ON: To enable building prometheus exporter.
-
Once build configuration is created, build the CMake targets - this includes building SDKs, and building unittests for API and SDK. Note that since API is header only library, no separate build is triggered for it.
$ cmake --build . --target all Scanning dependencies of target timestamp_test [ 0%] Building CXX object api/test/core/CMakeFiles/timestamp_test.dir/timestamp_test.cc.o [ 1%] Linking CXX executable timestamp_test ... Scanning dependencies of target w3c_tracecontext_test [ 99%] Building CXX object ext/test/w3c_tracecontext_test/CMakeFiles/w3c_tracecontext_test.dir/main.cc.o [100%] Linking CXX executable w3c_tracecontext_test [100%] Built target w3c_tracecontext_test $ -
Once CMake tests are built, run them with
ctestcommand$ ctest Test project /tmp/opentelemetry-cpp/build Start 1: trace.SystemTimestampTest.Construction ... Start 380: ext.http.urlparser.UrlParserTests.BasicTests ... 100% tests passed, 0 tests failed out of 380 $ -
Optionally install the header files for API, and generated targets and header files for SDK at custom/default install location.
$ cmake --install . --config Debug --prefix /<install_root>/ -- Installing: /<install-root>/lib/cmake/opentelemetry-cpp/opentelemetry-cpp-config.cmake -- Installing: /<install-root>/lib/cmake/opentelemetry-cpp/opentelemetry-cpp-config-version.cmake ... -- Installing: /<install-root>/include/opentelemetry//ext/zpages/static/tracez_index.h -- Installing: /<install-root>/include/opentelemetry//ext/zpages/static/tracez_style.h -- Installing: /<install-root>/include/opentelemetry//ext/zpages/threadsafe_span_data.h -- Installing: /<install-root>/lib/libopentelemetry_zpages.a $
Incorporating into an existing CMake Project
To use the library from a CMake project, you can locate it directly with
find_package and use the imported targets from generated package
configurations. As of now, this will import targets for both API and SDK. In
future, there may be separate packages for API and SDK which can be installed
and imported separtely according to need.
# CMakeLists.txt
find_package(opentelemetry-cpp CONFIG REQUIRED)
...
target_include_directories(foo PRIVATE ${OPENTELEMETRY_CPP_INCLUDE_DIRS})
target_link_libraries(foo PRIVATE ${OPENTELEMETRY_CPP_LIBRARIES})
Build instructions using Bazel
NOTE: Experimental, and not supported for all the components. Make sure the GoogleTest installation may fail if there is a different version of googletest already installed in system-defined path.
Prerequisites for Bazel
- A supported platform (e.g. Windows, macOS or Linux). Refer to Platforms Supported for more information.
- A compatible C++ compiler supporting at least C++11. Major compilers are supported. Refer to Supported Compilers for more information.
- Git for fetching opentelemetry-cpp source code from repository. To install Git, consult the Set up Git guide on GitHub.
- Bazel for building opentelemetry-cpp API, SDK with their unittests. We use 3.7.2 in our build system.
To install Bazel, consult the Installing Bazel guide.
Building as standalone Bazel Project
-
Getting the opentelementry-cpp source:
# Change to the directory where you want to create the code repository $ cd ~ $ mkdir source && cd source $ git clone https://github.com/open-telemetry/opentelemetry-cpp Cloning into 'opentelemetry-cpp'... ... Resolving deltas: 100% (3225/3225), done. $ -
Navigate to the repository cloned above, download the dependencies and build the source code:
$ cd opentelemtry-cpp $ bazel build //... bazel build -- //... -//exporters/otlp/... -//exporters/prometheus/... Extracting Bazel installation... Starting local Bazel server and connecting to it... INFO: Analyzed 121 targets (98 packages loaded, 3815 targets configured). INFO: Found 121 targets... INFO: From Compiling sdk/src/trace/tracer_context.cc: ... -
Once Bazel tests are built, run them with
bazel test //...command$ bazel test //... .. $ -
The build artifacts will be located under
bazel-bin
Incorporating into an existing Bazel Project
- WORKSPACE file:
http_archive(
name = "io_opentelemetry_cpp",
sha256 = "<sha256>",
strip_prefix = "opentelemetry-cpp-1.0.1",
urls = [
"https://github.com/open-telemetry/opentelemetry-cpp/archive/refs/tags/v1.0.1.tar.gz"
],
)
# Load OpenTelemetry dependencies after load.
load("@io_opentelemetry_cpp//bazel:repository.bzl", "opentelemetry_cpp_deps")
opentelemetry_cpp_deps()
# Load gRPC dependencies after load.
load("@com_github_grpc_grpc//bazel:grpc_deps.bzl", "grpc_deps")
grpc_deps()
# Load extra gRPC dependencies due to https://github.com/grpc/grpc/issues/20511
load("@com_github_grpc_grpc//bazel:grpc_extra_deps.bzl", "grpc_extra_deps")
grpc_extra_deps()
- Component level BUILD file:
cc_library(
name = "<name>"
...
deps = [
"@io_opentelemetry_cpp//api",
"@io_opentelemetry_cpp//exporters/otlp:otlp_exporter",
"@io_opentelemetry_cpp//sdk/src/trace",
...
],
...
)
Building shared libs for Windows
Windows DLL build is not supported. There are some constraints on how C++ DLLs work on Windows, specifically we can't safely allocate memory in one DLL and free it in another. For now, OpenTelemetry C++ targets need to be statically linked into the Windows applications.
Using Package Managers
If you are using Conan to manage your dependencies, add
opentelemetry-cpp/x.y.z to your
conanfile's requires, where x.y.z is the release version you want to use.
Please file issues here
if you experience problems with the packages.
If you are using vcpkg on your project
for external dependencies, then you can install the opentelemetry-cpp
package
with vcpkg install opentelemetry-cpp and follow the then displayed
descriptions. Please see the vcpkg project for any issues regarding the
packaging.
Please note, these packages are not officially provided and maintained by OpenTelemetry C++ project, and are just listed here to consolidate all such efforts for ease of developers.