datadog 2.41.0 → 2.43.0

data/lib/datadog/profiling/collectors/cpu_and_wall_time_worker.rb

@@ -33,6 +33,7 @@ module Datadog
         # @rbs allocation_counting_enabled: bool
         # @rbs gvl_profiling_enabled: bool
         # @rbs sighandler_sampling_enabled: bool
+        # @rbs waiting_for_gvl_threshold_ns: ::Integer
         # @rbs skip_idle_samples_for_testing: false
         # @rbs return: void
         def initialize(
@@ -45,6 +46,7 @@ module Datadog
           gvl_profiling_enabled:,
           sighandler_sampling_enabled:,
           cpu_sampling_interval_ms:,
+          waiting_for_gvl_threshold_ns:,
           # **NOTE**: This should only be used for testing; disabling the dynamic sampling rate will increase the
           # profiler overhead!
           dynamic_sampling_rate_enabled: true,
@@ -78,6 +80,7 @@ module Datadog
             sighandler_sampling_enabled: sighandler_sampling_enabled,
             skip_idle_samples_for_testing: skip_idle_samples_for_testing,
             cpu_sampling_interval_ms: cpu_sampling_interval_ms,
+            waiting_for_gvl_threshold_ns: waiting_for_gvl_threshold_ns,
           )
           @worker_thread = nil
           @failure_exception = nil

data/lib/datadog/profiling/collectors/info.rb

@@ -1,8 +1,7 @@
 # frozen_string_literal: true
 
-require "set"
-require "time"
 require "libdatadog"
+require_relative "../../core/utils/time"
 
 module Datadog
   module Profiling
@@ -71,10 +70,7 @@ module Datadog
         # Instead of trying to figure out real process start time by checking
         # /proc or some other complex/non-portable way, approximate start time
         # by time of requirement of this file.
-        #
-        # Note: this does not use Core::Utils::Time.now because this constant
-        # gets initialized before a user has a chance to configure the library.
-        START_TIME = Time.now.utc.freeze
+        START_TIME = Datadog::Core::Utils::Time.now.utc.freeze
 
         #: () -> ::Hash[::Symbol, untyped]
         def collect_platform_info

data/lib/datadog/profiling/collectors/thread_context.rb

@@ -19,7 +19,6 @@ module Datadog
           max_frames:,
           tracer:,
           endpoint_collection_enabled:,
-          waiting_for_gvl_threshold_ns:,
           otel_context_enabled:,
           native_filenames_enabled:,
           show_classes:
@@ -31,7 +30,6 @@ module Datadog
             max_frames: max_frames,
             tracer_context_key: tracer_context_key,
             endpoint_collection_enabled: endpoint_collection_enabled,
-            waiting_for_gvl_threshold_ns: waiting_for_gvl_threshold_ns,
             otel_context_enabled: otel_context_enabled,
             native_filenames_enabled: native_filenames_enabled,
             show_classes: show_classes,
@@ -44,7 +42,6 @@ module Datadog
           max_frames: 400,
           tracer: nil,
           endpoint_collection_enabled: false,
-          waiting_for_gvl_threshold_ns: 10_000_000,
           otel_context_enabled: false,
           native_filenames_enabled: true,
           show_classes: false,
@@ -56,7 +53,6 @@ module Datadog
             max_frames: max_frames,
             tracer: tracer,
             endpoint_collection_enabled: endpoint_collection_enabled,
-            waiting_for_gvl_threshold_ns: waiting_for_gvl_threshold_ns,
             otel_context_enabled: otel_context_enabled,
             native_filenames_enabled: native_filenames_enabled,
             show_classes: show_classes,

data/lib/datadog/profiling/component.rb

@@ -65,6 +65,7 @@ module Datadog
           gvl_profiling_enabled: enable_gvl_profiling?(settings, logger),
           sighandler_sampling_enabled: settings.profiling.advanced.sighandler_sampling_enabled,
           cpu_sampling_interval_ms: cpu_sampling_interval_ms,
+          waiting_for_gvl_threshold_ns: settings.profiling.advanced.waiting_for_gvl_threshold_ns,
         )
 
         internal_metadata = {
@@ -102,7 +103,6 @@ module Datadog
           max_frames: settings.profiling.advanced.max_frames,
           tracer: optional_tracer,
           endpoint_collection_enabled: settings.profiling.advanced.endpoint.collection.enabled,
-          waiting_for_gvl_threshold_ns: settings.profiling.advanced.waiting_for_gvl_threshold_ns,
           otel_context_enabled: settings.profiling.advanced.preview_otel_context_enabled,
           native_filenames_enabled: settings.profiling.advanced.native_filenames_enabled,
           show_classes: settings.profiling.advanced.experimental_show_classes_enabled,
@@ -227,11 +227,13 @@ module Datadog
           return false
         end
 
-        # Heap profiling relies on `ObjectSpace._id2ref`, which was removed on Ruby 4.1
-        # (https://bugs.ruby-lang.org/issues/22135).
-        if RubyVersion.is?(">= 4.1")
+        # Heap profiling tracks live objects using an `ObjectSpace::WeakMap`, which could corrupt its internal
+        # state during compaction before these versions (https://bugs.ruby-lang.org/issues/19529).
+        if RubyVersion.is?("< 3.1.4") || RubyVersion.is?(">= 3.2", "< 3.2.3")
           logger.warn(
-            "Heap profiling is currently incompatible with Ruby 4.1+ and has been disabled."
+            "Current Ruby version (#{RUBY_VERSION}) cannot support heap profiling due to a VM bug. " \
+            "Please upgrade to Ruby >= 3.1.4 or >= 3.2.3 in order to use this feature. " \
+            "Heap profiling has been disabled."
           )
           return false
         end
@@ -252,17 +254,7 @@ module Datadog
       private_class_method def self.enable_heap_size_profiling?(settings, heap_profiling_enabled, logger)
         heap_size_profiling_enabled = settings.profiling.advanced.experimental_heap_size_enabled
 
-        return false unless heap_profiling_enabled && heap_size_profiling_enabled
-
-        if RubyVersion.is?(">= 4")
-          logger.info(
-            "Heap live size profiling is currently incompatible with Ruby 4 and has been disabled. " \
-            "Heap live objects is not affected and remains enabled."
-          )
-          return false
-        end
-
-        true
+        heap_profiling_enabled && heap_size_profiling_enabled
       end
 
       private_class_method def self.no_signals_workaround_enabled?(settings, logger) # rubocop:disable Metrics/MethodLength

data/lib/datadog/tracing/contrib/active_record/events/sql.rb

@@ -46,6 +46,7 @@ module Datadog
               span.resource = payload.fetch(:sql)
               span.type = Tracing::Metadata::Ext::SQL::TYPE
 
+              span.set_tag(Tracing::Metadata::Ext::TAG_KIND, Tracing::Metadata::Ext::SpanKind::TAG_CLIENT)
               span.set_tag(Tracing::Metadata::Ext::TAG_COMPONENT, Ext::TAG_COMPONENT)
               span.set_tag(Tracing::Metadata::Ext::TAG_OPERATION, Ext::TAG_OPERATION_SQL)
 

data/lib/datadog/tracing/contrib/sequel/jdbc_connection_string.rb

@@ -0,0 +1,140 @@
+# frozen_string_literal: true
+
+module Datadog
+  module Tracing
+    module Contrib
+      module Sequel
+        # Extracts connection metadata from JDBC connection strings.
+        module JDBCConnectionString
+          MAX_BYTES = 8_192
+
+          # Matches JDBC connection strings whose subname uses `[transport:]//host...`.
+          CONNECTION_STRING_PATTERN =
+            %r{\Ajdbc:[a-z][a-z0-9+.-]*:(?:[a-z][a-z0-9+.-]*:)?//(?<subname>.+)\z}im
+
+          # Extracts the host and port from a string.
+          # The host can be a multi-host value, a bracketed IPv6 host, or an ordinary host.
+          HOST_AND_PORT_PATTERN =
+            /\A(?:\[(?<ipv6_host>[^\[\]]+)\]|(?<host>[^,]+(?:,[^,]+)+|[^:]+))(?::(?<port>\d+))?\z/
+
+          DATABASE_PROPERTY_PATTERN =
+            /(?:\A|[&;])(?:databaseName|database)=(?<value>[^&;]+)/i
+          LIBRARIES_PROPERTY_PATTERN =
+            /(?:\A|[&;])libraries=,*(?<value>[^,&;]+)/i
+          DATABASE_PATH_DELIMITER_PATTERN = %r{[:@\[\]?&=#]}
+
+          private_constant :MAX_BYTES, :CONNECTION_STRING_PATTERN,
+            :DATABASE_PROPERTY_PATTERN, :LIBRARIES_PROPERTY_PATTERN,
+            :DATABASE_PATH_DELIMITER_PATTERN, :HOST_AND_PORT_PATTERN
+
+          class << self
+            # The returned Hash guarantees the existence of all keys,
+            # but Hash values can be `nil` when not parseable from the connection string.
+            #
+            # @param connection_string [String, nil] the JDBC connection string to parse
+            # @return [Hash{Symbol => String, nil}]
+            #   - `:host` — host value when the subname uses authority syntax: `//host[:port]`
+            #   - `:port` — port when the subname uses authority syntax: `//host[:port]`
+            #   - `:database` — best-effort database name
+            def parse(connection_string)
+              # @type var result: metadata
+              result = {host: nil, port: nil, database: nil}
+              return result unless connection_string.is_a?(String) && connection_string.valid_encoding?
+
+              if connection_string.bytesize > MAX_BYTES
+                # Strip userinfo before truncation. Otherwise, an `@` beyond the limit could be
+                # discarded and leave a credential prefix looking like a valid authority/host.
+                authority_marker = connection_string.index("//")
+                if authority_marker
+                  authority_start = authority_marker + 2
+                  authority_end = [
+                    connection_string.index("/", authority_start),
+                    connection_string.index(";", authority_start),
+                    connection_string.index("?", authority_start),
+                  ].compact.min || connection_string.length
+                  userinfo_end = connection_string.rindex("@", authority_end - 1)
+
+                  if userinfo_end && userinfo_end >= authority_start
+                    connection_string =
+                      # Steep: https://github.com/soutaro/steep/issues/1219
+                      connection_string[0...authority_start] + # steep:ignore NoMethod
+                      connection_string[(userinfo_end + 1)..-1].to_s
+                  end
+                end
+
+                # Keep one extra byte, then let `chop` safely remove multi-byte unicode characters.
+                if connection_string.bytesize > MAX_BYTES
+                  connection_string = connection_string.byteslice(0, MAX_BYTES + 1).chop
+                end
+              end
+
+              match = CONNECTION_STRING_PATTERN.match(connection_string)
+              return result unless match
+
+              # We start with: `host[:port][/database][;properties][?query]`.
+              subname = match[:subname]
+
+              # Extract `query` from the end, leaving `host[:port][/database][;properties]`.
+              subname, query_separator, query = subname.partition("?")
+              query = nil if query_separator.empty?
+
+              # Extract `properties` next, leaving `host[:port][/database]`.
+              subname, properties_separator, properties = subname.partition(";")
+              properties = nil if properties_separator.empty?
+
+              # Separate the authority (`host[:port]`) from the optional database path.
+              authority, path = subname.split("/", 2)
+              return result if authority.nil? || authority.empty?
+
+              host, port = host_and_port_from_authority(authority)
+              return result unless host
+
+              database = database_from_path(path) ||
+                database_from_properties(properties) || database_from_properties(query)
+
+              {host: host, port: port, database: database}
+            rescue Encoding::CompatibilityError, ArgumentError
+              result
+            end
+
+            private
+
+            def database_from_path(path)
+              return if path.nil? || path.empty?
+
+              database = path.split(DATABASE_PATH_DELIMITER_PATTERN, 2).first
+              return if database.nil? || database.empty?
+
+              database
+            end
+
+            def host_and_port_from_authority(authority)
+              # Discard optional `userinfo@`, retaining only `host[:port]`.
+              authority = authority.rpartition("@").last
+              return if authority.empty?
+
+              match = HOST_AND_PORT_PATTERN.match(authority)
+              return unless match
+
+              # Only one of `ipv6_host` or `host` will be populated.
+              host = match[:ipv6_host] || match[:host]
+              port = match[:port]
+
+              [host, port]
+            end
+
+            def database_from_properties(properties)
+              return unless properties
+
+              database = DATABASE_PROPERTY_PATTERN.match(properties)
+              return database[:value] if database
+
+              libraries = LIBRARIES_PROPERTY_PATTERN.match(properties)
+              libraries && libraries[:value]
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/lib/datadog/tracing/contrib/sequel/utils.rb

@@ -1,10 +1,9 @@
 # frozen_string_literal: true
 
-require "uri"
-
 require_relative "../../metadata/ext"
 require_relative "../utils/database"
 require_relative "ext"
+require_relative "jdbc_connection_string"
 require_relative "../ext"
 require_relative "../span_attribute_schema"
 
@@ -14,11 +13,6 @@ module Datadog
       module Sequel
         # General purpose functions for Sequel
         module Utils
-          JDBC_URI_PATTERN = %r{\Ajdbc:(?<vendor>[a-z][a-z0-9+.-]*):(?<location>//[^\r\n]*)\z}i
-          DATABASE_PROPERTY_PATTERN =
-            /(?:\A|[&;])(?<key>databaseName|database|libraries)=(?<value>[^&;]+)/i
-          private_constant :JDBC_URI_PATTERN, :DATABASE_PROPERTY_PATTERN
-
           class << self
             # Ruby database connector library
             #
@@ -42,34 +36,6 @@ module Datadog
               Contrib::Utils::Database.normalize_vendor(database.database_type.to_s)
             end
 
-            # Parses URI-style JDBC connection strings, extracting host, port, and
-            # (best-effort) database name. Unsupported or ambiguous forms return empty
-            # metadata rather than potentially incorrect tags.
-            def parse_jdbc_uri(uri)
-              result = {host: nil, port: nil, database: nil}
-              return result unless uri.is_a?(String) && uri.valid_encoding?
-
-              match = JDBC_URI_PATTERN.match(uri)
-              return result unless match
-
-              vendor = match[:vendor].downcase
-              location, properties = match[:location].split(";", 2)
-
-              # Several JDBC vendors append properties with semicolons, outside the URI
-              # grammar. Parse the URI-compatible location separately from those properties.
-              parsed = URI.parse("#{vendor}:#{location}")
-
-              host = parsed.hostname
-              port = parsed.port
-
-              database = database_from_path(parsed.path) ||
-                database_from_properties(properties) || database_from_properties(parsed.query)
-
-              {host: host, port: port&.to_s, database: database}
-            rescue URI::InvalidURIError, Encoding::CompatibilityError, ArgumentError
-              result
-            end
-
             def parse_opts(sql, opts, db_opts, dataset = nil)
               # Prepared statements don't provide their sql query in the +sql+ parameter.
               if !sql.is_a?(String) && dataset&.respond_to?(:prepared_sql) &&
@@ -121,25 +87,38 @@ module Datadog
               Contrib::Analytics.set_sample_rate(span, analytics_sample_rate) if analytics_enabled?
             end
 
-            # Resolves the connection host/port/database for a Sequel::Database. When the
-            # connection string is a JDBC URL (Sequel's JDBC adapter, used on JRuby), the
-            # host/port/database are parsed from it regardless of whether opts[:host] is set.
+            # Resolves the connection host/port/database for a Sequel::Database. For Sequel's
+            # JDBC adapter (used on JRuby), metadata is parsed from the JDBC connection string
+            # regardless of whether opts[:host] is set.
             def connection_metadata(db)
               opts = db.opts || {}
               host = opts[:host]
               port = opts[:port]
               database = opts[:database]
 
-              # A JDBC URL (in :uri, :url, or :database) can carry credentials, so always parse
-              # it and emit only the parsed database name -- never the raw connection string.
-              conn = opts[:uri] || opts[:url] || opts[:database]
-              is_jdbc = conn.is_a?(String) && conn.byteslice(0, 5)&.casecmp("jdbc:") == 0
+              # A JDBC connection string (in :uri, :url, or :database) can carry credentials, so
+              # always emit only parsed metadata -- never the raw connection string.
+              connection_string = opts[:uri] || opts[:url] || opts[:database]
+              is_jdbc = connection_string.is_a?(String) &&
+                connection_string.byteslice(0, 5)&.casecmp("jdbc:") == 0
               if is_jdbc
-                parsed = parse_jdbc_uri(conn)
+                parsed = JDBCConnectionString.parse(connection_string)
+
+                # JNDI/DataSource-managed connections keep only a lookup name in opts (e.g.
+                # "jdbc:jndi:..."), so nothing can be parsed from it. Recover the endpoint from the
+                # live connection's JDBC metadata, the same way Sequel resolves JNDI. This stays a
+                # fallback rather than the primary source: it requires a connection checkout and
+                # some drivers report no connection string, whereas the opts value is free and
+                # already present for direct connections (and non-JDBC adapters have no such
+                # metadata at all).
+                if parsed[:host].nil? && parsed[:port].nil? && parsed[:database].nil?
+                  fallback_metadata = jdbc_metadata_from_connection(db)
+                  parsed = fallback_metadata if fallback_metadata
+                end
 
-                # Sequel's JDBC adapter connects with the URL and ignores separate
+                # Sequel's JDBC adapter connects with the connection string and ignores separate
                 # :host/:port options, unlike native adapters where those options take precedence.
-                if !parsed[:host].nil? || !parsed[:port].nil? || !parsed[:database].nil?
+                if parsed[:host] || parsed[:port] || parsed[:database]
                   host = parsed[:host]
                   port = parsed[:port]
                 end
@@ -151,24 +130,32 @@ module Datadog
 
             private
 
-            def database_from_path(path)
-              return unless path&.start_with?("/")
-
-              database = path[1..-1]
-              return if database.empty? || database.include?("/")
-
-              database
-            end
-
-            def database_from_properties(properties)
-              return unless properties
-
-              match = DATABASE_PROPERTY_PATTERN.match(properties)
-              return unless match
+            # Resolves host/port/database from the live connection's JDBC metadata
+            # (java.sql.DatabaseMetaData#getURL), for JNDI/DataSource connections whose opts hold
+            # only a lookup name. Returns the parsed metadata, or nil when it can't be resolved.
+            #
+            # Only *parsed, credential-free* metadata is memoized -- the raw connection string is
+            # used transiently and never stored or logged. A completed lookup that yields no usable
+            # connection string is a permanent property of the connection, so it is cached to avoid
+            # re-checking out a connection on every query. A raised error is treated as transient
+            # (pool checkout timeout, dropped connection, ...) and left uncached, so a later query
+            # can retry once connectivity recovers.
+            def jdbc_metadata_from_connection(db)
+              return db.instance_variable_get(:@datadog_jdbc_metadata) if db.instance_variable_defined?(:@datadog_jdbc_metadata)
+
+              connection_string =
+                begin
+                  db.synchronize do |conn|
+                    conn.get_meta_data.get_url if conn.respond_to?(:get_meta_data)
+                  end
+                rescue => e
+                  Datadog.logger.debug { "Sequel: unable to resolve JDBC connection metadata (#{e.class})" }
+                  return nil
+                end
 
-              database = match[:value]
-              database = database.split(",", 2).first if match[:key].casecmp("libraries").zero?
-              database
+              metadata = connection_string && JDBCConnectionString.parse(connection_string)
+              db.instance_variable_set(:@datadog_jdbc_metadata, metadata)
+              metadata
             end
 
             def datadog_configuration

data/lib/datadog/tracing/distributed/baggage.rb

@@ -3,7 +3,6 @@
 require_relative "../ext"
 require_relative "../metadata/ext"
 require_relative "../trace_digest"
-require_relative "datadog_tags_codec"
 require_relative "../utils"
 require_relative "helpers"
 require "uri"

data/lib/datadog/tracing/distributed/datadog.rb

@@ -2,7 +2,7 @@
 
 require_relative "../metadata/ext"
 require_relative "../trace_digest"
-require_relative "datadog_tags_codec"
+require_relative "trace_state/datadog"
 require_relative "../utils"
 require_relative "helpers"
 
@@ -129,7 +129,7 @@ module Datadog
         def inject_tags!(tags, data)
           return set_tags_propagation_error(reason: "disabled") if tags_disabled?
 
-          encoded_tags = DatadogTagsCodec.encode(tags)
+          encoded_tags = TraceState::Datadog.encode(tags)
 
           return set_tags_propagation_error(reason: "inject_max_size") if tags_too_large?(encoded_tags, scenario: "inject")
 
@@ -155,7 +155,7 @@ module Datadog
           return set_tags_propagation_error(reason: "disabled") if tags_disabled?
           return set_tags_propagation_error(reason: "extract_max_size") if tags_too_large?(tags, scenario: "extract")
 
-          tags_hash = DatadogTagsCodec.decode(tags)
+          tags_hash = TraceState::Datadog.decode(tags)
           # Only extract keys with the expected Datadog prefix
           tags_hash.select! do |key, _|
             key.start_with?(Tracing::Metadata::Ext::Distributed::TAGS_PREFIX) && key != EXCLUDED_TAG

data/lib/datadog/tracing/distributed/propagation.rb

@@ -146,6 +146,9 @@ module Datadog
                 span_id: parent_id,
                 trace_state: tracecontext_digest.trace_state,
                 trace_state_unknown_fields: tracecontext_digest.trace_state_unknown_fields,
+                trace_otel_random_value: tracecontext_digest.trace_otel_random_value,
+                trace_otel_threshold: tracecontext_digest.trace_otel_threshold,
+                trace_otel_unknown_fields: tracecontext_digest.trace_otel_unknown_fields,
                 trace_distributed_tags: distributed_tags
               )
             end