datadog 2.7.1 → 2.18.0

data/lib/datadog/core/configuration/settings.rb

@@ -18,7 +18,7 @@ module Datadog
     module Configuration
       # Global configuration settings for the Datadog library.
       # @public_api
-      # rubocop:disable Metrics/BlockLength
+      # standard:disable Metrics/BlockLength
       class Settings
         include Base
 
@@ -123,7 +123,7 @@ module Datadog
           # @return [Boolean]
           option :debug do |o|
             o.env [Datadog::Core::Configuration::Ext::Diagnostics::ENV_DEBUG_ENABLED,
-                   Datadog::Core::Configuration::Ext::Diagnostics::ENV_OTEL_LOG_LEVEL]
+              Datadog::Core::Configuration::Ext::Diagnostics::ENV_OTEL_LOG_LEVEL]
             o.default false
             o.type :bool
             o.env_parser do |value|
@@ -137,7 +137,7 @@ module Datadog
             o.after_set do |enabled|
               # Enable rich debug print statements.
               # We do not need to unnecessarily load 'pp' unless in debugging mode.
-              require 'pp' if enabled
+              require 'pp' if enabled # standard:disable Lint/RedundantRequireStatement
             end
           end
 
@@ -313,7 +313,7 @@ module Datadog
 
             # Can be used to enable/disable the collection of heap profiles.
             #
-            # This feature is alpha and disabled by default
+            # This feature is in preview and disabled by default. Requires Ruby 3.1+.
             #
             # @warn To enable heap profiling you are required to also enable allocation profiling.
             #
@@ -326,12 +326,12 @@ module Datadog
 
             # Can be used to enable/disable the collection of heap size profiles.
             #
-            # This feature is alpha and enabled by default when heap profiling is enabled.
+            # This feature is in preview and by default is enabled whenever heap profiling is enabled.
             #
-            # @warn To enable heap size profiling you are required to also enable allocation and heap profiling.
+            # @warn Heap size profiling depends on allocation and heap profiling, so they must be enabled as well.
             #
-            # @default `DD_PROFILING_EXPERIMENTAL_HEAP_SIZE_ENABLED` environment variable as a boolean, otherwise
-            # whatever the value of DD_PROFILING_EXPERIMENTAL_HEAP_ENABLED is.
+            # @default `DD_PROFILING_EXPERIMENTAL_HEAP_SIZE_ENABLED` environment variable as a boolean, otherwise it
+            # follows the value of `experimental_heap_enabled`.
             option :experimental_heap_size_enabled do |o|
               o.type :bool
               o.env 'DD_PROFILING_EXPERIMENTAL_HEAP_SIZE_ENABLED'
@@ -341,17 +341,19 @@ module Datadog
             # Can be used to configure the heap sampling rate: a heap sample will be collected for every x allocation
             # samples.
             #
-            # The lower the value, the more accuracy in heap tracking but the bigger the overhead. In particular, a
-            # value of 1 will track ALL allocations samples for heap profiles.
+            # The higher the value, the less accuracy in heap tracking but the smaller the overhead.
+            #
+            # If you needed to tweak this, please tell us why on <https://github.com/DataDog/dd-trace-rb/issues/new>,
+            # so we can fix it!
             #
             # The effective heap sampling rate in terms of allocations (not allocation samples) can be calculated via
             # effective_heap_sample_rate = allocation_sample_rate * heap_sample_rate.
             #
-            # @default `DD_PROFILING_EXPERIMENTAL_HEAP_SAMPLE_RATE` environment variable, otherwise `10`.
+            # @default `DD_PROFILING_EXPERIMENTAL_HEAP_SAMPLE_RATE` environment variable, otherwise `1`.
             option :experimental_heap_sample_rate do |o|
               o.type :int
               o.env 'DD_PROFILING_EXPERIMENTAL_HEAP_SAMPLE_RATE'
-              o.default 10
+              o.default 1
             end
 
             # Can be used to disable checking which version of `libmysqlclient` is being used by the `mysql2` gem.
@@ -454,22 +456,38 @@ module Datadog
               o.after_set do |_, _, precedence|
                 unless precedence == Datadog::Core::Configuration::Option::Precedence::DEFAULT
                   Core.log_deprecation(key: :experimental_crash_tracking_enabled) do
-                    'The profiling.advanced.experimental_crash_tracking_enabled setting has been deprecated for removal '\
+                    'The profiling.advanced.experimental_crash_tracking_enabled setting has been deprecated for removal ' \
                     'and no longer does anything. Please remove it from your Datadog.configure block.'
                   end
                 end
               end
             end
 
-            # Enables GVL profiling. This will show when threads are waiting for GVL in the timeline view.
-            #
-            # This is a preview feature and disabled by default. It requires Ruby 3.2+.
-            #
-            # @default `DD_PROFILING_PREVIEW_GVL_ENABLED` environment variable as a boolean, otherwise `false`
+            # @deprecated Use {:gvl_enabled} instead.
             option :preview_gvl_enabled do |o|
               o.type :bool
-              o.env 'DD_PROFILING_PREVIEW_GVL_ENABLED'
               o.default false
+              o.after_set do |_, _, precedence|
+                unless precedence == Datadog::Core::Configuration::Option::Precedence::DEFAULT
+                  Datadog.logger.warn(
+                    'The profiling.advanced.preview_gvl_enabled setting has been deprecated for removal and ' \
+                    'no longer does anything. Please remove it from your Datadog.configure block. ' \
+                    'GVL profiling is now controlled by the profiling.advanced.gvl_enabled setting instead.'
+                  )
+                end
+              end
+            end
+
+            # Controls GVL profiling. This will show when threads are waiting for GVL in the timeline view.
+            #
+            # This feature requires Ruby 3.2+.
+            #
+            # @default `DD_PROFILING_GVL_ENABLED` environment variable as a boolean, otherwise `true`
+            option :gvl_enabled do |o|
+              o.type :bool
+              o.deprecated_env 'DD_PROFILING_PREVIEW_GVL_ENABLED'
+              o.env 'DD_PROFILING_GVL_ENABLED'
+              o.default true
             end
 
             # Controls the smallest time period the profiler will report a thread waiting for the GVL.
@@ -526,6 +544,16 @@ module Datadog
               o.env 'DD_PROFILING_HEAP_CLEAN_AFTER_GC_ENABLED'
               o.default true
             end
+
+            # Controls if the profiler should use native filenames for frames in stack traces for functions implemented using
+            # native code. Setting to `false` will make the profiler fall back to default Ruby stack trace behavior (only show .rb files).
+            #
+            # @default true
+            option :native_filenames_enabled do |o|
+              o.type :bool
+              o.env 'DD_PROFILING_NATIVE_FILENAMES_ENABLED'
+              o.default true
+            end
           end
 
           # @public_api
@@ -554,6 +582,12 @@ module Datadog
             o.type :bool
           end
 
+          option :experimental_runtime_id_enabled do |o|
+            o.type :bool
+            o.env ['DD_TRACE_EXPERIMENTAL_RUNTIME_ID_ENABLED', 'DD_RUNTIME_METRICS_RUNTIME_ID_ENABLED']
+            o.default false
+          end
+
           option :opts, default: {}, type: :hash
           option :statsd
         end
@@ -603,38 +637,31 @@ module Datadog
           o.type :hash, nilable: true
           o.env [Core::Environment::Ext::ENV_TAGS, Core::Environment::Ext::ENV_OTEL_RESOURCE_ATTRIBUTES]
           o.env_parser do |env_value|
-            values = if env_value.include?(',')
-                       env_value.split(',')
-                     else
-                       env_value.split(' ') # rubocop:disable Style/RedundantArgument
-                     end
-            values.map! do |v|
-              v.gsub!(/\A[\s,]*|[\s,]*\Z/, '')
-
-              v.empty? ? nil : v
-            end
-
-            values.compact!
-            values.each_with_object({}) do |tag, tags|
-              key, value = tag.split(':', 2)
-              if value.nil?
-                # support tags/attributes delimited by the OpenTelemetry separator (`=`)
-                key, value = tag.split('=', 2)
-              end
-              next if value.nil? || value.empty?
-
-              # maps OpenTelemetry semantic attributes to Datadog tags
-              case key.downcase
-              when 'deployment.environment'
-                tags['env'] = value
-              when 'service.version'
-                tags['version'] = value
-              when 'service.name'
-                tags['service'] = value
-              else
-                tags[key] = value
+            # Parses a string containing key-value pairs and returns a hash.
+            # Key-value pairs are delimited by ':' OR `=`, and pairs are separated by whitespace, comma, OR BOTH.
+            result = {}
+            unless env_value.nil? || env_value.empty?
+              # falling back to comma as separator
+              sep = env_value.include?(',') ? ',' : ' '
+              # split by separator
+              env_value.split(sep).each do |tag|
+                tag.strip!
+                next if tag.empty?
+
+                # tag by : or = (for OpenTelemetry)
+                key, val = tag.split(/[:=]/, 2).map(&:strip)
+                val ||= ''
+                # maps OpenTelemetry semantic attributes to Datadog tags
+                key = case key.downcase
+                when 'deployment.environment' then 'env'
+                when 'service.version' then 'version'
+                when 'service.name' then 'service'
+                else key
+                end
+                result[key] = val unless key.empty?
               end
             end
+            result
           end
           o.setter do |new_value, old_value|
             raw_tags = new_value || {}
@@ -863,6 +890,26 @@ module Datadog
             o.type :float
             o.default 1.0
           end
+
+          # Enable log collection for telemetry. Log collection only works when telemetry is enabled and
+          # logs are enabled.
+          # @default `DD_TELEMETRY_LOG_COLLECTION_ENABLED` environment variable, otherwise `true`.
+          # @return [Boolean]
+          option :log_collection_enabled do |o|
+            o.type :bool
+            o.env Core::Telemetry::Ext::ENV_LOG_COLLECTION
+            o.default true
+          end
+
+          # For internal use only.
+          # Enables telemetry debugging through the Datadog platform.
+          #
+          # @default `false`.
+          # @return [Boolean]
+          option :debug do |o|
+            o.type :bool
+            o.default false
+          end
         end
 
         # Remote configuration
@@ -932,11 +979,35 @@ module Datadog
           end
         end
 
+        # Tracer specific configuration starting with APM (e.g. DD_APM_TRACING_ENABLED).
+        # @public_api
+        settings :apm do
+          # Tracing as a transport
+          # @public_api
+          settings :tracing do
+            # Enables tracing as transport.
+            # Disabling it will set sampling priority to -1 (FORCE_DROP) on most traces,
+            # (which tells to the agent to drop these traces)
+            # except heartbeat ones (1 per minute) and manually kept ones (sampling priority to 2) (e.g. appsec events)
+            #
+            # This is different than `DD_TRACE_ENABLED`, which completely disables tracing (sends no trace at all),
+            # while this will send heartbeat traces (1 per minute) so that the service is considered alive in the backend.
+            #
+            # @default `DD_APM_TRACING_ENABLED` environment variable, otherwise `true`
+            # @return [Boolean]
+            option :enabled do |o|
+              o.env Configuration::Ext::APM::ENV_TRACING_ENABLED
+              o.default true
+              o.type :bool
+            end
+          end
+        end
+
         # TODO: Tracing should manage its own settings.
         #       Keep this extension here for now to keep things working.
         extend Datadog::Tracing::Configuration::Settings
       end
-      # rubocop:enable Metrics/BlockLength
+      # standard:enable Metrics/BlockLength
     end
   end
 end

data/lib/datadog/core/configuration/stable_config.rb

@@ -0,0 +1,22 @@
+# frozen_string_literal: true
+
+module Datadog
+  module Core
+    module Configuration
+      # Import config from config files (fleet automation)
+      module StableConfig
+        def self.extract_configuration
+          if (libdatadog_api_failure = Datadog::Core::LIBDATADOG_API_FAILURE)
+            Datadog.config_init_logger.debug("Cannot enable stable config: #{libdatadog_api_failure}")
+            return {}
+          end
+          Configurator.new.get
+        end
+
+        def self.configuration
+          @configuration ||= StableConfig.extract_configuration
+        end
+      end
+    end
+  end
+end

data/lib/datadog/core/configuration.rb

@@ -172,6 +172,10 @@ module Datadog
         end
       end
 
+      def config_init_logger
+        configuration? ? logger : logger_without_configuration
+      end
+
       # Gracefully shuts down all components.
       #
       # Components will still respond to method calls as usual,
@@ -231,19 +235,21 @@ module Datadog
         end
 
         COMPONENTS_WRITE_LOCK.synchronize do
-          begin
-            yield write_components
-          rescue ThreadError => e
-            logger_without_components.error(
-              'Detected deadlock during datadog initialization. ' \
-              'Please report this at https://github.com/DataDog/dd-trace-rb/blob/master/CONTRIBUTING.md#found-a-bug' \
-              "\n\tSource:\n\t#{Array(e.backtrace).join("\n\t")}"
-            )
-            nil
-          end
+          yield write_components
+        rescue ThreadError => e
+          logger_without_components.error(
+            'Detected deadlock during datadog initialization. ' \
+            'Please report this at https://github.com/datadog/dd-trace-rb/blob/master/CONTRIBUTING.md#found-a-bug' \
+            "\n\tSource:\n\t#{Array(e.backtrace).join("\n\t")}"
+          )
+          nil
         end
       end
 
+      def configuration?
+        (defined?(@configuration) && @configuration) != nil
+      end
+
       def components?
         # This does not need to grab the COMPONENTS_READ_LOCK because it's not returning the components
         (defined?(@components) && @components) != nil
@@ -258,14 +264,25 @@ module Datadog
       def replace_components!(settings, old)
         components = Components.new(settings)
 
+        # Carry over state from existing components to the new ones.
+        # Currently:
+        # 1. If we already started the remote component (which
+        #    happens after a request goes through installed Rack middleware),
+        #    we will start the new remote component as well.
+        # 2. If telemetry has been enabled and is enabled in the new
+        #    component tree, we send AppConfigurationChange event instead
+        #    of AppStarted event.
+        old_state = old.state
+
         old.shutdown!(components)
-        components.startup!(settings)
+        components.startup!(settings, old_state: old_state)
         components
       end
 
       def logger_without_components
         # Use default logger without initializing components.
         # This enables logging during initialization, otherwise we'd run into deadlocks.
+
         @temp_logger ||= begin
           logger = configuration.logger.instance || Core::Logger.new($stdout)
           logger.level = configuration.diagnostics.debug ? ::Logger::DEBUG : configuration.logger.level
@@ -273,6 +290,21 @@ module Datadog
         end
       end
 
+      def logger_without_configuration
+        # There's rare cases where we need to use logger during configuration initialization,
+        # such as reading stable config. In this case we cannot access configuration.
+
+        @temp_config_logger ||= begin
+          debug_env_value = ENV[Ext::Diagnostics::ENV_DEBUG_ENABLED]&.strip&.downcase
+          debug_value = debug_env_value == 'true' || debug_env_value == '1'
+
+          logger = Core::Logger.new($stdout)
+          # We cannot access config and the default level is INFO, so we need to set the level manually
+          logger.level = debug_value ? ::Logger::DEBUG : ::Logger::INFO
+          logger
+        end
+      end
+
       # Called from our at_exit hook whenever there was a pending Interrupt exception (e.g. typically due to ctrl+c)
       # to print a nice message whenever we're taking a bit longer than usual to finish the process.
       def handle_interrupt_shutdown!

data/lib/datadog/{tracing → core}/contrib/rails/utils.rb

@@ -1,9 +1,7 @@
 # frozen_string_literal: true
 
-require_relative '../analytics'
-
 module Datadog
-  module Tracing
+  module Core
     module Contrib
       module Rails
         # common utilities for Rails

data/lib/datadog/core/crashtracking/component.rb

@@ -3,7 +3,6 @@
 require 'libdatadog'
 
 require_relative 'tag_builder'
-require_relative 'agent_base_url'
 require_relative '../utils/only_once'
 require_relative '../utils/at_fork_monkey_patch'
 
@@ -19,20 +18,11 @@ module Datadog
       #
       # Methods prefixed with _native_ are implemented in `crashtracker.c`
       class Component
-        LIBDATADOG_API_FAILURE =
-          begin
-            require "libdatadog_api.#{RUBY_VERSION[/\d+.\d+/]}_#{RUBY_PLATFORM}"
-            nil
-          rescue LoadError => e
-            e.message
-          end
-
         ONLY_ONCE = Core::Utils::OnlyOnce.new
 
         def self.build(settings, agent_settings, logger:)
           tags = TagBuilder.call(settings)
-          agent_base_url = AgentBaseUrl.resolve(agent_settings)
-          logger.warn('Missing agent base URL; cannot enable crash tracking') unless agent_base_url
+          agent_base_url = agent_settings.url
 
           ld_library_path = ::Libdatadog.ld_library_path
           logger.warn('Missing ld_library_path; cannot enable crash tracking') unless ld_library_path
@@ -73,7 +63,7 @@ module Datadog
               # Must NOT reference `self` here, as only the first instance will
               # be captured by the ONLY_ONCE and we want to pick the latest active one
               # (which may have different tags or agent config)
-              Datadog.send(:components).crashtracker&.update_on_fork
+              Datadog.send(:components, allow_initialization: false)&.crashtracker&.update_on_fork
             end
           end
         end
@@ -102,7 +92,8 @@ module Datadog
             path_to_crashtracking_receiver_binary: path_to_crashtracking_receiver_binary,
             ld_library_path: ld_library_path,
             tags_as_array: tags.to_a,
-            upload_timeout_seconds: 1
+            # @ivoanjo: On my machine this needs to be > 5 seconds, and seems to work with 10; the extra 15 is extra margin
+            upload_timeout_seconds: 15,
           )
           logger.debug("Crash tracking action: #{action} successful")
         rescue => e

data/lib/datadog/core/crashtracking/tag_builder.rb

@@ -1,9 +1,7 @@
 # frozen_string_literal: true
 
+require_relative '../tag_builder'
 require_relative '../utils'
-require_relative '../environment/socket'
-require_relative '../environment/identity'
-require_relative '../environment/git'
 
 module Datadog
   module Core
@@ -11,27 +9,11 @@ module Datadog
       # This module builds a hash of tags
       module TagBuilder
         def self.call(settings)
-          hash = {
-            'host' => Environment::Socket.hostname,
-            'process_id' => Process.pid.to_s,
-            'runtime_engine' => Environment::Identity.lang_engine,
-            'runtime-id' => Environment::Identity.id,
-            'runtime_platform' => Environment::Identity.lang_platform,
-            'runtime_version' => Environment::Identity.lang_version,
-            'env' => settings.env,
-            'service' => settings.service,
-            'version' => settings.version,
-            'git.repository_url' => Environment::Git.git_repository_url,
-            'git.commit.sha' => Environment::Git.git_commit_sha,
+          hash = Core::TagBuilder.tags(settings).merge(
             'is_crash' => 'true',
-            'language' => 'ruby',
-            'library_version' => Core::Environment::Identity.gem_datadog_version,
-          }.compact
+          )
 
-          # Make sure everything is an utf-8 string, to avoid encoding issues in downstream
-          settings.tags.merge(hash).each_with_object({}) do |(key, value), h|
-            h[Utils.utf8_encode(key)] = Utils.utf8_encode(value)
-          end
+          Utils.encode_tags(hash)
         end
       end
     end

data/lib/datadog/core/diagnostics/environment_logger.rb

@@ -79,7 +79,7 @@ module Datadog
 
           # @return [String] current time in ISO8601 format
           def date
-            Time.now.utc.iso8601
+            Core::Utils::Time.now.utc.iso8601
           end
 
           # Best portable guess of OS information.

data/lib/datadog/core/encoding.rb

@@ -10,6 +10,7 @@ module Datadog
       # Encoder interface that provides the logic to encode traces and service
       # @abstract
       module Encoder
+        # :nocov:
         def content_type
           raise NotImplementedError
         end
@@ -23,6 +24,13 @@ module Datadog
         def encode(_)
           raise NotImplementedError
         end
+
+        # Deserializes a value serialized with {#encode}.
+        # This method is used for debugging purposes.
+        def decode(_)
+          raise NotImplementedError
+        end
+        # :nocov:
       end
 
       # Encoder for the JSON format
@@ -41,8 +49,12 @@ module Datadog
           JSON.dump(obj)
         end
 
+        def decode(obj)
+          JSON.parse(obj)
+        end
+
         def join(encoded_data)
-          "[#{encoded_data.join(',')}]"
+          "[#{encoded_data.join(",")}]"
         end
       end
 
@@ -62,6 +74,10 @@ module Datadog
           MessagePack.pack(obj)
         end
 
+        def decode(obj)
+          MessagePack.unpack(obj)
+        end
+
         def join(encoded_data)
           packer = MessagePack::Packer.new
           packer.write_array_header(encoded_data.size)

data/lib/datadog/core/environment/agent_info.rb

@@ -0,0 +1,78 @@
+# frozen_string_literal: true
+
+module Datadog
+  module Core
+    module Environment
+      # Retrieves the agent's `/info` endpoint data.
+      # This data can be used to determine the capabilities of the local Datadog agent.
+      #
+      # @example Example response payload
+      #   {
+      #     "version" : "7.57.2",
+      #     "git_commit" : "38ba0c7",
+      #     "endpoints" : [ "/v0.4/traces", "/v0.4/services", "/v0.7/traces", "/v0.7/config" ],
+      #     "client_drop_p0s" : true,
+      #     "span_meta_structs" : true,
+      #     "long_running_spans" : true,
+      #     "evp_proxy_allowed_headers" : [ "Content-Type", "Accept-Encoding", "Content-Encoding", "User-Agent" ],
+      #     "config" : {
+      #       "default_env" : "none",
+      #       "target_tps" : 10,
+      #       "max_eps" : 200,
+      #       "receiver_port" : 8126,
+      #       "receiver_socket" : "/var/run/datadog/apm.socket",
+      #       "connection_limit" : 0,
+      #       "receiver_timeout" : 0,
+      #       "max_request_bytes" : 26214400,
+      #       "statsd_port" : 8125,
+      #       "analyzed_spans_by_service" : { },
+      #       "obfuscation" : {
+      #         "elastic_search" : true,
+      #         "mongo" : true,
+      #         "sql_exec_plan" : false,
+      #         "sql_exec_plan_normalize" : false,
+      #         "http" : {
+      #           "remove_query_string" : false,
+      #           "remove_path_digits" : false
+      #         },
+      #         "remove_stack_traces" : false,
+      #         "redis" : {
+      #           "Enabled" : true,
+      #           "RemoveAllArgs" : false
+      #         },
+      #         "memcached" : {
+      #           "Enabled" : true,
+      #           "KeepCommand" : false
+      #         }
+      #       }
+      #     },
+      #     "peer_tags" : null
+      #   }
+      #
+      # @see https://github.com/DataDog/datadog-agent/blob/f07df0a3c1fca0c83b5a15f553bd994091b0c8ac/pkg/trace/api/info.go#L20
+      class AgentInfo
+        attr_reader :agent_settings, :logger
+
+        def initialize(agent_settings, logger: Datadog.logger)
+          @agent_settings = agent_settings
+          @logger = logger
+          @client = Remote::Transport::HTTP.root(agent_settings: agent_settings, logger: logger)
+        end
+
+        # Fetches the information from the agent.
+        # @return [Datadog::Core::Remote::Transport::HTTP::Negotiation::Response] the response from the agent
+        # @return [nil] if an error occurred while fetching the information
+        def fetch
+          res = @client.send_info
+          return unless res.ok?
+
+          res
+        end
+
+        def ==(other)
+          other.is_a?(self.class) && other.agent_settings == agent_settings
+        end
+      end
+    end
+  end
+end

data/lib/datadog/core/environment/cgroup.rb

@@ -23,20 +23,18 @@ module Datadog
 
         def descriptors(process = 'self')
           [].tap do |descriptors|
-            begin
-              filepath = "/proc/#{process}/cgroup"
-
-              if File.exist?(filepath)
-                File.foreach("/proc/#{process}/cgroup") do |line|
-                  line = line.strip
-                  descriptors << parse(line) unless line.empty?
-                end
+            filepath = "/proc/#{process}/cgroup"
+
+            if File.exist?(filepath)
+              File.foreach("/proc/#{process}/cgroup") do |line|
+                line = line.strip
+                descriptors << parse(line) unless line.empty?
               end
-            rescue StandardError => e
-              Datadog.logger.error(
-                "Error while parsing cgroup. Cause: #{e.class.name} #{e.message} Location: #{Array(e.backtrace).first}"
-              )
             end
+          rescue => e
+            Datadog.logger.error(
+              "Error while parsing cgroup. Cause: #{e.class.name} #{e.message} Location: #{Array(e.backtrace).first}"
+            )
           end
         end