Module: Datadog::Core::Configuration

Defined in:
lib/datadog/core/configuration.rb,
lib/datadog/core/configuration/ext.rb,
lib/datadog/core/configuration/base.rb,
lib/datadog/core/configuration/option.rb,
lib/datadog/core/configuration/options.rb,
lib/datadog/core/configuration/settings.rb,
lib/datadog/core/configuration/components.rb,
lib/datadog/core/configuration/option_definition.rb,
lib/datadog/core/configuration/agent_settings_resolver.rb

Overview

Configuration provides a unique access point for configurations

Defined Under Namespace

Modules: Base, Ext, Options Classes: Option, OptionDefinition, Settings

Instance Attribute Summary collapse

Instance Method Summary collapse

Instance Attribute Details

#configuration ⇒ Datadog::Core::Configuration::Settings

Current Datadog configuration.

Access to non-global configuration will raise an error.

To modify the configuration, use #configure.

Returns:



49
50
51
 
# File 'lib/datadog/core/configuration.rb', line 49

def configuration
  @configuration ||= Settings.new
end

Instance Method Details

#configuration_for(target, option = nil) ⇒ Object

Get configuration changes applied only to a specific Ruby object, via #configure_onto. An example of an object with specific configuration:

client = Net::HTTP.new(host, port)
Datadog.configure_onto(client, service_name: 'api-requests', split_by_domain: true)
config = Datadog.configuration_for(client)
config[:service_name] # => 'api-requests'
config[:split_by_domain] # => true

Parameters:

  • target (Object) —

    the object to receive configuration options

  • option (Object) (defaults to: nil) —

    an option to retrieve from the object configuration



147
148
149
150
151
152
 
# File 'lib/datadog/core/configuration.rb', line 147

def configuration_for(target, option = nil)
  pin = Pin.get_from(target)
  return pin unless option

  pin[option] if pin
end

#configure {|c| ... } ⇒ Object

Apply global configuration changes to Datadog. An example of a #configure call:

Datadog.configure do |c|
  c.service = 'my-service'
  c.env = 'staging'
  # c.diagnostics.debug = true # Enables debug output
end

See Settings for all available options, defaults, and available environment variables for configuration.

Only permits access to global configuration settings; others will raise an error. If you wish to configure a setting for a specific Datadog component (e.g. Tracing), use the corresponding Datadog::COMPONENT.configure method instead.

Because many configuration changes require restarting internal components, invoking #configure is the only safe way to change Datadog configuration.

Successive calls to #configure maintain the previous configuration values: configuration is additive between #configure calls.

The yielded configuration c comes pre-populated from environment variables, if any are applicable.

Yield Parameters:



80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
 
# File 'lib/datadog/core/configuration.rb', line 80

def configure
  configuration = self.configuration
  yield(configuration)

  safely_synchronize do |write_components|
    write_components.call(
      if components?
        replace_components!(configuration, @components)
      else
        components = build_components(configuration)
        components.telemetry.started!
        components
      end
    )
  end

  configuration
end

#configure_onto(target, **opts) ⇒ Object

Apply configuration changes only to a specific Ruby object.

Certain integrations or Datadog features may use these settings to customize behavior for this object.

An example of a #configure_onto call:

client = Net::HTTP.new(host, port)
Datadog.configure_onto(client, service_name: 'api-requests', split_by_domain: true)

In this example, it will configure the client object with custom options service_name: 'api-requests', split_by_domain: true. The Net::HTTP integration will then use these customized options when the client is used, whereas other clients will use the service_name: 'http-requests' configuration provided to the Datadog.configure call block.

#configure_onto is used to separate cases where spans generated by certain objects require exceptional options.

The configuration keyword arguments provided should match well known options defined in the integration or feature that would use them.

For example, for Datadog.configure_onto(redis_client, **opts), opts can be any of the options in the Redis Tracing::Contrib::Redis::Configuration::Settings class.

Parameters:

  • target (Object) —

    the object to receive configuration options

  • opts (Hash) —

    keyword arguments respective to the integration this object belongs to



129
130
131
 
# File 'lib/datadog/core/configuration.rb', line 129

def configure_onto(target, **opts)
  Pin.set_on(target, **opts)
end

#health_metrics ⇒ Object

Internal Statsd metrics collection.

The list of metrics collected can be found in Diagnostics::Ext::Health::Metrics.



158
159
160
 
# File 'lib/datadog/core/configuration.rb', line 158

def health_metrics
  components.health_metrics
end

#logger ⇒ Object 



162
163
164
165
166
167
168
169
170
171
172
 
# File 'lib/datadog/core/configuration.rb', line 162

def logger
  # avoid initializing components if they didn't already exist
  current_components = components(allow_initialization: false)

  if current_components
    @temp_logger = nil
    current_components.logger
  else
    logger_without_components
  end
end

#shutdown! ⇒ Object

Gracefully shuts down all components.

Components will still respond to method calls as usual, but might not internally perform their work after shutdown.

This avoids errors being raised across the host application during shutdown, while allowing for graceful decommission of resources.

Components won't be automatically reinitialized after a shutdown.



183
184
185
186
187
 
# File 'lib/datadog/core/configuration.rb', line 183

def shutdown!
  safely_synchronize do
    @components.shutdown! if components?
  end
end