Skip to content

optimizely/nginx-datadog

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

datadog tracing nginx

Datadog NGINX Module

This repository contains the source code for the ngx_http_datadog_module, an NGINX module that integrates Datadog APM and Application Security Management into NGINX.

Usage

  1. Download a gzipped tarball from a recent release, extract it to wherever nginx looks for modules (e.g. /usr/lib/nginx/modules/).
  2. Add the following line to the top of the main nginx configuration (e.g. /etc/nginx/nginx.conf):
load_module modules/ngx_http_datadog_module.so;

Tracing is automatically added to all endpoints by default. For more information, see the API documentation.

Compatibility

Important

We provide support for NGINX versions up to their End Of Life, extended by one year. Aligned with the NGINX release cycle, this entails support for the four most recent NGINX versions.

If you plan to add tracing features to an older NGINX version using our module, please check out the build section for guidance.

There are two tarballs (the actual executable module and, separately, the debug symbols) per each combination of: 1) nginx version, 2) architecture, 3) whether AppSec is built in or not. The main tarball contains a single file, ngx_http_datadog_module.so, which is the Datadog nginx module.

The naming convention is:

  • ngx_http_datadog_module-<arch>-<version>.so.tgz for builds without appsec support and
  • ngx_http_datadog_module-appsec-<arch>-<version>.so.tgz for builds with appsec support.

Important

The AppSec variants require nginx to have been built with --threads (thread support).

Supported architectures (<arch>) are amd64 and arm64.

Default Behavior

Unless otherwise configured, ngx_http_datadog_module adds the following default behavior to nginx:

Tracing

  • Connect to the Datadog agent at http://localhost:8126.
  • Create one span per request:
    • Service name is "nginx".
    • Operation name is "nginx.request".
    • Resource name is "$request_method $uri", e.g. "GET /api/book/0-345-24223-8/title".
    • Includes multiple http.* tags.

Custom configuration can be specified via the datadog_* family of directives in nginx's configuration file, or via environment variables.

Enabling AppSec

To enable AppSec, besides using the correct binary (the relase artifact with "-appsec") in the name, it's necessary to edit the nginx configuration:

  • Set datadog_appsec_enabled on;.
  • Define one (or more thread pools).
  • Choose which thread pool AppSec will use, either on a global or a per-location basis.

For more information, see the documentation.

Build

Requirements:

  • Recent C and C++ toolchain (clang or gcc/g++) (must support at least some C++20 features).
  • CMake v3.24 or newer.
  • Architecture must be x86_64 or arm64.

For enhanced usability, we provide a GNU make compatible Makefile.

NGINX_VERSION=1.25.2 make build

You can set the environment variable WAF to ON to build an AppSec-supporting module:

WAF=ON NGINX_VERSION=1.25.2 make build

The resulting nginx module is .build/ngx\_http\_datadog\_module.so

The build target does the following:

  • Download a source release of nginx based on the NGINX\_VERSION environment variable.
  • Initialize the source tree of dd-trace-cpp as a git submodule.
  • Build dd-trace-cpp and the Datadog nginx module together using CMake.

make clean deletes CMake's build directory. make clobber deletes everything done by the build.

Build in Docker

make build-musl

The build-musl target builds against musl and libc++ a glibc-compatible module. The Dockerfile for the docker image used in the process can be found in build_env/Dockerfile.

Test

See test/README.md.

Acknowledgements

This project is based largely on previous work. See CREDITS.md.

About

nginx module for distributed tracing

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • C++ 49.5%
  • Python 34.0%
  • C 4.3%
  • Shell 3.8%
  • JavaScript 2.6%
  • CMake 2.4%
  • Other 3.4%