datadog 2.20.0 → 2.26.0

data/lib/datadog/core/remote/worker.rb

@@ -23,51 +23,47 @@ module Datadog
         attr_reader :logger
 
         def start
-          logger.debug { 'remote worker starting' }
+          logger.debug { "remote worker starting (pid: #{Process.pid})" }
 
-          acquire_lock
+          @mutex.synchronize do
+            if @stopped
+              logger.debug('remote worker: refusing to restart after previous stop')
+              return
+            end
 
-          if @stopped
-            logger.debug('remote worker: refusing to restart after previous stop')
-            return
-          end
-
-          return if @starting || @started
+            return if @starting || @started
 
-          @starting = true
+            @starting = true
 
-          thread = Thread.new { poll(@interval) }
-          thread.name = self.class.name
-          thread.thread_variable_set(:fork_safe, true)
-          @thr = thread
+            thread = Thread.new { poll(@interval) }
+            thread.name = self.class.name
+            thread.thread_variable_set(:fork_safe, true)
+            @thr = thread
 
-          @started = true
-          @starting = false
+            @started = true
+            @starting = false
+          end
 
           logger.debug { 'remote worker started' }
-        ensure
-          release_lock
         end
 
         def stop
-          logger.debug { 'remote worker stopping' }
+          logger.debug { "remote worker stopping (pid: #{Process.pid})" }
 
-          acquire_lock
+          @mutex.synchronize do
+            thread = @thr
 
-          thread = @thr
+            if thread
+              thread.kill
+              thread.join
+            end
 
-          if thread
-            thread.kill
-            thread.join
+            @started = false
+            @thr = nil
+            @stopped = true
           end
 
-          @started = false
-          @thr = nil
-          @stopped = true
-
           logger.debug { 'remote worker stopped' }
-        ensure
-          release_lock
         end
 
         def started?
@@ -76,14 +72,6 @@ module Datadog
 
         private
 
-        def acquire_lock
-          @mutex.lock
-        end
-
-        def release_lock
-          @mutex.unlock
-        end
-
         def poll(interval)
           loop do
             break unless @mutex.synchronize { @starting || @started }

data/lib/datadog/core/runtime/ext.rb

@@ -13,7 +13,6 @@ module Datadog
         # @public_api
         module Metrics
           ENV_ENABLED = 'DD_RUNTIME_METRICS_ENABLED'
-          ENV_OTEL_METRICS_EXPORTER = 'OTEL_METRICS_EXPORTER'
 
           METRIC_CLASS_COUNT = 'runtime.ruby.class_count'
           METRIC_GC_PREFIX = 'runtime.ruby.gc'

data/lib/datadog/core/runtime/metrics.rb

@@ -8,6 +8,7 @@ require_relative '../environment/gc'
 require_relative '../environment/thread_count'
 require_relative '../environment/vm_cache'
 require_relative '../environment/yjit'
+require_relative '../environment/process'
 
 module Datadog
   module Core
@@ -24,6 +25,9 @@ module Datadog
 
           # Initialize the collection of runtime-id
           @runtime_id_enabled = options.fetch(:experimental_runtime_id_enabled, false)
+
+          # Initialized process tags support
+          @process_tags_enabled = options.fetch(:experimental_propagate_process_tags_enabled, false)
         end
 
         # Associate service with runtime metrics
@@ -111,6 +115,11 @@ module Datadog
 
             # Add runtime-id dynamically because it might change during runtime.
             options[:tags].concat(["runtime-id:#{Core::Environment::Identity.id}"]) if @runtime_id_enabled
+
+            # Add process tags when enabled
+            if @process_tags_enabled
+              options[:tags].concat(Core::Environment::Process.tags)
+            end
           end
         end
 
@@ -119,7 +128,8 @@ module Datadog
         attr_reader \
           :service_tags,
           :services,
-          :runtime_id_enabled
+          :runtime_id_enabled,
+          :process_tags_enabled
 
         def compile_service_tags!
           @service_tags = services.to_a.collect do |service|

data/lib/datadog/core/semaphore.rb

@@ -20,10 +20,7 @@ module Datadog
 
       def wait(timeout = nil)
         wake_lock.synchronize do
-          # steep specifies that the second argument to wait is of type
-          # ::Time::_Timeout which for some reason is not Numeric and is not
-          # castable from Numeric.
-          wake.wait(wake_lock, timeout) # steep:ignore
+          wake.wait(wake_lock, timeout)
         end
       end
 

data/lib/datadog/core/tag_builder.rb

@@ -8,10 +8,6 @@ module Datadog
   module Core
     # This module builds a hash of tags.
     #
-    # When changing or adding the tags, make sure they are kept in sync with
-    # https://docs.google.com/spreadsheets/d/1LOGMf4c4Avbtn36uZ2SWvhIGKRPLM1BoWkUP4JYj7hA/
-    # (Datadog internal link).
-    #
     # @api private
     module TagBuilder
       def self.fixed_environment_tags

data/lib/datadog/core/tag_normalizer.rb

@@ -0,0 +1,84 @@
+# frozen_string_literal: true
+
+require_relative 'utils'
+
+module Datadog
+  module Core
+    # @api private
+    module TagNormalizer
+      # Normalization logic used for tag keys and values that the Trace Agent has for traces
+      # Useful for ensuring that tag keys and values are normalized consistently
+      # An use case for now is Process Tags which need to be sent across various intakes (profiling, tracing, etc.) consistently
+
+      module_function
+
+      INVALID_TAG_CHARACTERS = %r{[^\p{L}0-9_\-:./]}
+      LEADING_INVALID_CHARS_NO_DIGITS = %r{\A[^\p{L}:]++}
+      LEADING_INVALID_CHARS_WITH_DIGITS = %r{\A[^\p{L}0-9:./]++}
+      MAX_BYTE_SIZE = 200 # Represents the general max tag length
+      MAX_PROCESS_VALUE_BYTE_SIZE = 100 # Represents the max tag length for process tags
+      VALID_ASCII_TAG = %r{\A[a-z:][a-z0-9:./-]*\z}
+
+      # Based on https://github.com/DataDog/datadog-agent/blob/45799c842bbd216bcda208737f9f11cade6fdd95/pkg/trace/traceutil/normalize.go#L131
+      # Specifically for general normalization:
+      # - Must be valid UTF-8
+      # - Invalid characters are replaced with an underscore
+      # - Leading non-letter characters are removed but colons are kept
+      # - Trailing non-letter characters are removed
+      # - Trailing underscores are removed
+      # - Consecutive underscores are merged into a single underscore
+      # - Maximum length is 200 characters
+      # If it's a tag value, allow it to start with a digit
+      # @param original_value [String] The original string
+      # @param remove_digit_start_char [Boolean] - whether to remove the leading digit (currently only used for tag values)
+      # @return [String] The normalized string
+      def self.normalize(original_value, remove_digit_start_char: false)
+        # DEV-3.0: Ideally this encode call should be replaced with Datadog::Core::Utils.utf8_encode once it
+        #          is safe to modify the default behavior.
+        value = original_value.to_s.encode('UTF-8', invalid: :replace, undef: :replace)
+        value.strip!
+        return "" if value.empty?
+
+        return value if value.bytesize <= MAX_BYTE_SIZE &&
+          value.match?(VALID_ASCII_TAG)
+
+        if value.bytesize > MAX_BYTE_SIZE
+          value = value.byteslice(0, MAX_BYTE_SIZE)
+          value.scrub!("")
+        end
+
+        value.downcase!
+        value.gsub!(INVALID_TAG_CHARACTERS, '_')
+
+        # The Trace Agent allows tag values to start with a number so this logic is here too
+        leading_invalid_regex = remove_digit_start_char ? LEADING_INVALID_CHARS_NO_DIGITS : LEADING_INVALID_CHARS_WITH_DIGITS
+        value.sub!(leading_invalid_regex, "")
+
+        value.squeeze!('_') if value.include?('__')
+        value.delete_suffix!('_')
+
+        value
+      end
+
+      # Process tags values follow an additional piece of normalization:
+      # - must not be more than 100 bytes
+      # - and must not contain colons
+      # @param value [String] The original string
+      # @return [String] The normalized string
+      def self.normalize_process_value(value)
+        value = normalize(value)
+        return value if value.empty?
+
+        value.tr!(':', '_')
+        value.squeeze!('_') if value.include?('__')
+
+        if value.bytesize > MAX_PROCESS_VALUE_BYTE_SIZE
+          value = value.byteslice(0, MAX_PROCESS_VALUE_BYTE_SIZE) || value
+          value.scrub!("")
+        end
+
+        value
+      end
+    end
+  end
+end

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

@@ -14,12 +14,26 @@ require_relative '../utils/forking'
 module Datadog
   module Core
     module Telemetry
-      # Telemetry entrypoint, coordinates sending telemetry events at various points in app lifecycle.
-      # Note: Telemetry does not spawn its worker thread in fork processes, thus no telemetry is sent in forked processes.
+      # Telemetry entry point, coordinates sending telemetry events at
+      # various points in application lifecycle.
       #
       # @api private
       class Component
-        attr_reader :enabled, :logger, :transport, :worker
+        ENDPOINT_COLLECTION_MESSAGE_LIMIT = 300
+
+        ONLY_ONCE = Utils::OnlyOnce.new
+
+        attr_reader :enabled
+        attr_reader :logger
+        attr_reader :transport
+        attr_reader :worker
+        attr_reader :settings
+        attr_reader :agent_settings
+        attr_reader :metrics_manager
+
+        # Alias for consistency with other components.
+        # TODO Remove +enabled+ method
+        alias_method :enabled?, :enabled
 
         include Core::Utils::Forking
         include Telemetry::Logging
@@ -48,6 +62,17 @@ module Datadog
           logger:,
           enabled:
         )
+          ONLY_ONCE.run do
+            Utils::AtForkMonkeyPatch.apply!
+
+            # All of the other at fork monkey patch callbacks reference
+            # globals, follow that pattern here to avoid having the component
+            # referenced via the at fork callbacks.
+            Datadog::Core::Utils::AtForkMonkeyPatch.at_fork(:child) do
+              Datadog.send(:components, allow_initialization: false)&.telemetry&.after_fork
+            end
+          end
+
           @enabled = enabled
           @log_collection_enabled = settings.telemetry.log_collection_enabled
           @logger = logger
@@ -107,13 +132,17 @@ module Datadog
           @worker&.enabled = false
         end
 
-        def start(initial_event_is_change = false)
-          return if !@enabled
+        def start(initial_event_is_change = false, components:)
+          return unless enabled?
 
           initial_event = if initial_event_is_change
-            Event::SynthAppClientConfigurationChange.new(agent_settings: @agent_settings)
+            Event::SynthAppClientConfigurationChange.new(
+              components: components,
+            )
           else
-            Event::AppStarted.new(agent_settings: @agent_settings)
+            Event::AppStarted.new(
+              components: components,
+            )
           end
 
           @worker.start(initial_event)
@@ -130,41 +159,50 @@ module Datadog
         end
 
         def emit_closing!
-          return if !@enabled || forked?
+          return unless enabled?
 
           @worker.enqueue(Event::AppClosing.new)
         end
 
         def integrations_change!
-          return if !@enabled || forked?
+          return unless enabled?
 
           @worker.enqueue(Event::AppIntegrationsChange.new)
         end
 
         def log!(event)
-          return if !@enabled || forked? || !@log_collection_enabled
+          return unless enabled? && @log_collection_enabled
 
           @worker.enqueue(event)
         end
 
         # Wait for the worker to send out all events that have already
         # been queued, up to 15 seconds. Returns whether all events have
-        # been flushed.
+        # been flushed, or nil if telemetry is disabled.
         #
         # @api private
-        def flush
-          return if !@enabled || forked?
+        def flush(timeout: nil)
+          return unless enabled?
 
-          @worker.flush
+          @worker.flush(timeout: timeout)
         end
 
         # Report configuration changes caused by Remote Configuration.
         def client_configuration_change!(changes)
-          return if !@enabled || forked?
+          return unless enabled?
 
           @worker.enqueue(Event::AppClientConfigurationChange.new(changes, 'remote_config'))
         end
 
+        # Report application endpoints
+        def app_endpoints_loaded(endpoints, page_size: ENDPOINT_COLLECTION_MESSAGE_LIMIT)
+          return unless enabled?
+
+          endpoints.each_slice(page_size).with_index do |endpoints_slice, i|
+            @worker.enqueue(Event::AppEndpointsLoaded.new(endpoints_slice, is_first: i.zero?))
+          end
+        end
+
         # Increments a count metric.
         def inc(namespace, metric_name, value, tags: {}, common: true)
           @metrics_manager.inc(namespace, metric_name, value, tags: tags, common: common)
@@ -189,6 +227,22 @@ module Datadog
         def distribution(namespace, metric_name, value, tags: {}, common: true)
           @metrics_manager.distribution(namespace, metric_name, value, tags: tags, common: common)
         end
+
+        # When a fork happens, we generally need to do two things inside the
+        # child proess:
+        # 1. Restart the worker.
+        # 2. Discard any events and metrics that were submitted in the
+        #    parent process (because they will be sent out in the parent
+        #    process, sending them in the child would cause duplicate
+        #    submission).
+        def after_fork
+          # We cannot simply create a new instance of metrics manager because
+          # it is referenced from other objects (e.g. the worker).
+          # We must reset the existing instance.
+          @metrics_manager.clear
+
+          worker&.send(:after_fork_monkey_patched)
+        end
       end
     end
   end

data/lib/datadog/core/telemetry/emitter.rb

@@ -31,15 +31,15 @@ module Datadog
           seq_id = self.class.sequence.next
           payload = Request.build_payload(event, seq_id, debug: debug?)
           res = @transport.send_telemetry(request_type: event.type, payload: payload)
-          logger.debug { "Telemetry sent for event `#{event.type}` (response code: #{res.code})" }
+          if res.ok?
+            logger.debug { "Telemetry sent for event `#{event.type}`" }
+          else
+            logger.debug { "Failed to send telemetry for event `#{event.type}`: #{res.inspect}" }
+          end
           res
         rescue => e
           logger.debug {
-            "Unable to send telemetry request for event `#{begin
-              event.type
-            rescue
-              "unknown"
-            end}`: #{e}"
+            "Unable to send telemetry request for event `#{event.respond_to?(:type) ? event.type : event.to_s}`: #{e.class}: #{e}"
           }
           Core::Transport::InternalErrorResponse.new(e)
         end

data/lib/datadog/core/telemetry/event/app_endpoints_loaded.rb

@@ -0,0 +1,30 @@
+# frozen_string_literal: true
+
+require_relative 'base'
+
+module Datadog
+  module Core
+    module Telemetry
+      module Event
+        # Telemetry event class for sending 'app-endpoints' payload
+        class AppEndpointsLoaded < Base
+          def initialize(endpoints, is_first:)
+            @endpoints = endpoints
+            @is_first = !!is_first
+          end
+
+          def type
+            'app-endpoints'
+          end
+
+          def payload
+            {
+              is_first: @is_first,
+              endpoints: @endpoints
+            }
+          end
+        end
+      end
+    end
+  end
+end