Using instrumentation libraries

当你开发应用时，可能会使用第三方库和框架来加快开发进度。如果你随后使用 OpenTelemetry 对应用进行插桩，你可能希望避免额外花时间为所用的第三方库和框架手动添加链路、日志和指标。

许多库和框架已经原生支持 OpenTelemetry，或者通过 OpenTelemetry 的插桩获得支持， 因此它们能够生成可导出到可观测性后端的遥测数据。

如果你正在为使用第三方库或框架的应用或服务进行插桩， 请按照以下说明学习如何为你的依赖项使用原生插桩库和插桩库。

使用原生插桩库

如果某个库默认就支持 OpenTelemetry，你只需在应用中添加并配置 OpenTelemetry SDK， 就可以获取该库发出的链路、指标和日志。

该库可能需要一些额外的插桩配置。请查阅该库的文档以了解更多信息。

• Elasticsearch Ruby Client

Use Instrumentation Libraries

If a library does not come with OpenTelemetry out of the box, you can use instrumentation libraries in order to generate telemetry data for a library or framework.

For example, if you are using Rails and enable opentelemetry-instrumentation-rails, your running Rails app will automatically generate telemetry data for inbound requests to your controllers.

Configuring all instrumentation libraries

OpenTelemetry Ruby provides the metapackage opentelemetry-instrumentation-all that bundles all ruby-based instrumentation libraries into a single package. It’s a convenient way to add telemetry for all your libraries with minimal effort:

gem 'opentelemetry-sdk'
gem 'opentelemetry-exporter-otlp'
gem 'opentelemetry-instrumentation-all'

and configure it early in your application lifecycle. See the example below using a Rails initializer:

# config/initializers/opentelemetry.rb
require 'opentelemetry/sdk'
require 'opentelemetry/exporter/otlp'
require 'opentelemetry/instrumentation/all'
OpenTelemetry::SDK.configure do |c|
  c.service_name = '<YOUR_SERVICE_NAME>'
  c.use_all() # enables all instrumentation!
end

This will install all instrumentation libraries and enable the ones that match up to libraries you’re using in your app.

Overriding configuration for specific instrumentation libraries

If you are enabling all instrumentation but want to override the configuration for a specific one, call use_all with a configuration map parameter, where the key represents the library, and the value is its specific configuration parameter.

For example, here’s how you can install all instrumentations except the Redis instrumentation into your app:

require 'opentelemetry/sdk'
require 'opentelemetry/instrumentation/all'

OpenTelemetry::SDK.configure do |c|
  config = {'OpenTelemetry::Instrumentation::Redis' => { enabled: false }}
  c.use_all(config)
end

To override more instrumentation, add another entry in the config map.

Overriding configuration for specific instrumentation libraries with environment variables

You can also disable specific instrumentation libraries using environment variables. An instrumentation disabled by an environment variable takes precedence over local config. The convention for environment variable names is the library name, upcased with :: replaced by underscores, OPENTELEMETRY shortened to OTEL_LANG, and _ENABLED appended.

For example, the environment variable name for OpenTelemetry::Instrumentation::Sinatra is OTEL_RUBY_INSTRUMENTATION_SINATRA_ENABLED.

export OTEL_RUBY_INSTRUMENTATION_SINATRA_ENABLED=false

Configuring specific instrumentation libraries

If you prefer more selectively installing and using only specific instrumentation libraries, you can do that too. For example, here’s how to use only Sinatra and Faraday, with Faraday being configured with an additional configuration parameter.

First, install the specific instrumentation libraries you know you want to use:

gem install opentelemetry-instrumentation-sinatra
gem install opentelemetry-instrumentation-faraday

Then configure them:

require 'opentelemetry/sdk'

# install all compatible instrumentation with default configuration
OpenTelemetry::SDK.configure do |c|
  c.use 'OpenTelemetry::Instrumentation::Sinatra'
  c.use 'OpenTelemetry::Instrumentation::Faraday', { opt: 'value' }
end

Configuring specific instrumentation libraries with environment variables

You can also define the option for specific instrumentation libraries using environment variables. By convention, the environment variable will be the name of the instrumentation, upcased with :: replaced by underscores, OPENTELEMETRY shortened to OTEL_{LANG}, and _CONFIG_OPTS appended.

For example, the environment variable name for OpenTelemetry::Instrumentation::Faraday is OTEL_RUBY_INSTRUMENTATION_FARADAY_CONFIG_OPTS. A value of peer_service=new_service;span_kind=client overrides the options set from previous section for Faraday.

export OTEL_RUBY_INSTRUMENTATION_FARADAY_CONFIG_OPTS="peer_service=new_service;span_kind=client"

The following table lists the acceptable format for values according to the option data type:

Next steps

Instrumentation libraries are the easiest way to generate lots of useful telemetry data about your Ruby apps. But they don’t generate data specific to your application’s logic! To do that, you’ll need to enrich the instrumentation from instrumentation libraries with your own instrumentation code.