datadog 2.37.0 → 2.39.0

data/lib/datadog/di/context.rb

@@ -14,6 +14,8 @@ module Datadog
         target_self: nil,
         path: nil, caller_locations: nil,
         serialized_entry_args: nil,
+        entry_capture_expressions: nil,
+        entry_capture_evaluation_errors: nil,
         return_value: nil, duration: nil, exception: nil)
         @probe = probe
         @settings = settings
@@ -23,6 +25,8 @@ module Datadog
         @path = path
         @caller_locations = caller_locations
         @serialized_entry_args = serialized_entry_args
+        @entry_capture_expressions = entry_capture_expressions
+        @entry_capture_evaluation_errors = entry_capture_evaluation_errors
         @return_value = return_value
         @duration = duration
         @exception = exception
@@ -44,6 +48,10 @@ module Datadog
 
       attr_reader :serialized_entry_args
 
+      attr_reader :entry_capture_expressions
+
+      attr_reader :entry_capture_evaluation_errors
+
       # Return value for the method, for a method probe
       attr_reader :return_value
 
@@ -55,9 +63,7 @@ module Datadog
 
       def serialized_locals
         # TODO cache?
-        locals && serializer.serialize_vars(locals,
-          depth: probe.max_capture_depth || settings.dynamic_instrumentation.max_capture_depth,
-          attribute_count: probe.max_capture_attribute_count || settings.dynamic_instrumentation.max_capture_attribute_count,)
+        locals && serializer.serialize_vars(locals, **probe.snapshot_serializer_limits(settings))
       end
 
       def fetch(var_name)

data/lib/datadog/di/contrib/railtie.rb

@@ -6,7 +6,7 @@ module Datadog
       # Railtie class initializes dynamic instrumentation contrib code
       # in Rails environments.
       class Railtie < Rails::Railtie
-        initializer 'datadog.dynamic_instrumentation.initialize' do |app|
+        initializer "datadog.dynamic_instrumentation.initialize" do |app|
           Contrib.load_now
         end
       end

data/lib/datadog/di/contrib.rb

@@ -1,14 +1,14 @@
 # frozen_string_literal: true
 
-require_relative '../core/contrib/rails/utils'
+require_relative "../core/contrib/rails/utils"
 
 module Datadog
   module DI
     module Contrib
       module_function def load_now_or_later
         if Datadog::Core::Contrib::Rails::Utils.railtie_supported?
-          Datadog.logger.debug('di: loading contrib/railtie')
-          require_relative 'contrib/railtie'
+          Datadog.logger.debug("di: loading contrib/railtie")
+          require_relative "contrib/railtie"
         else
           load_now
         end
@@ -19,8 +19,8 @@ module Datadog
       # dependencies are loaded (or potentially loaded).
       module_function def load_now
         if defined?(ActiveRecord::Base)
-          Datadog.logger.debug('di: loading contrib/active_record')
-          require_relative 'contrib/active_record'
+          Datadog.logger.debug("di: loading contrib/active_record")
+          require_relative "contrib/active_record"
         end
       end
     end

data/lib/datadog/di/el/compiler.rb

@@ -24,12 +24,12 @@ module Datadog
 
         # Steep: https://github.com/soutaro/steep/issues/363
         OPERATORS = { # steep:ignore IncompatibleAssignment
-          'eq' => '==',
-          'ne' => '!=',
-          'ge' => '>=',
-          'gt' => '>',
-          'le' => '<=',
-          'lt' => '<',
+          "eq" => "==",
+          "ne" => "!=",
+          "ge" => ">=",
+          "gt" => ">",
+          "le" => "<=",
+          "lt" => "<",
         }.freeze
 
         # Steep: https://github.com/soutaro/steep/issues/363
@@ -45,8 +45,8 @@ module Datadog
 
         # Steep: https://github.com/soutaro/steep/issues/363
         MULTI_ARG_METHODS = { # steep:ignore IncompatibleAssignment
-          'and' => '&&',
-          'or' => '||',
+          "and" => "&&",
+          "or" => "||",
         }.freeze
 
         def compile_partial(ast)
@@ -57,29 +57,29 @@ module Datadog
             end
             op, target = ast.first
             case op
-            when 'ref'
+            when "ref"
               unless String === target
                 raise DI::Error::InvalidExpression, "Bad ref value type: #{target.class}: #{target}"
               end
               case target
-              when '@it'
-                'current_item'
-              when '@key'
-                'current_key'
-              when '@value'
-                'current_value'
-              when '@return'
+              when "@it"
+                "current_item"
+              when "@key"
+                "current_key"
+              when "@value"
+                "current_value"
+              when "@return"
                 # For @return, @duration and @exception we shadow
                 # instance variables.
                 "context.return_value"
-              when '@duration'
+              when "@duration"
                 # There is no way to explicitly format the duration.
                 # TODO come up with better formatting?
                 # We could format to a string here but what if customer
                 # has @duration as part of an expression and wants
                 # to retain it as a number?
-                "(context.duration * 1000)"
-              when '@exception'
+                "(context.duration && context.duration * 1000)"
+              when "@exception"
                 "context.exception"
               else
                 # Ruby technically allows all kinds of symbols in variable
@@ -89,7 +89,7 @@ module Datadog
                 unless target =~ %r{\A(@?)([a-zA-Z0-9_]+)\z}
                   raise DI::Error::BadVariableName, "Bad variable name: #{target}"
                 end
-                method_name = (($1 == '@') ? 'iref' : 'ref')
+                method_name = (($1 == "@") ? "iref" : "ref")
                 "#{method_name}('#{target}')"
               end
             when *SINGLE_ARG_METHODS
@@ -111,12 +111,12 @@ module Datadog
               end
               compiled_op = MULTI_ARG_METHODS[op]
               "(#{compiled_targets.join(" #{compiled_op} ")})"
-            when 'substring'
+            when "substring"
               unless Array === target && target.length == 3
                 raise DI::Error::InvalidExpression, "Improper #{op} syntax"
               end
               "#{op}(#{target.map { |arg| "(#{compile_partial(arg)})" }.join(", ")})"
-            when 'not'
+            when "not"
               "!(#{compile_partial(target)})"
             when *OPERATORS.keys
               unless Array === target && target.length == 2
@@ -125,7 +125,7 @@ module Datadog
               first, second = target
               operator = OPERATORS.fetch(op)
               "(#{compile_partial(first)}) #{operator} (#{compile_partial(second)})"
-            when 'any', 'all', 'filter'
+            when "any", "all", "filter"
               "#{op}(#{compile_partial(target.first)}) { |current_item, current_key, current_value| #{compile_partial(target.last)} }"
             else
               raise DI::Error::InvalidExpression, "Unknown operation: #{op}"
@@ -151,16 +151,16 @@ module Datadog
         # WARNING: the result of this method is included in eval'd code,
         # it must be sanitized to avoid injection.
         def var_name_maybe(target)
-          if Hash === target && target.length == 1 && target.keys.first == 'ref' &&
+          if Hash === target && target.length == 1 && target.keys.first == "ref" &&
               String === (value = target.values.first)
             escape(value)
           else
-            '(expression)'
+            "(expression)"
           end
         end
 
         def escape(needle)
-          needle.gsub("\\") { "\\\\" }.gsub('"') { "\\\"" }.gsub('#') { "\\#" }
+          needle.gsub("\\") { "\\\\" }.gsub('"') { "\\\"" }.gsub("#") { "\\#" }
         end
       end
     end

data/lib/datadog/di/el.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
-require_relative 'el/expression'
-require_relative 'el/compiler'
-require_relative 'el/evaluator'
+require_relative "el/expression"
+require_relative "el/compiler"
+require_relative "el/evaluator"

data/lib/datadog/di/instrumenter.rb

@@ -1,8 +1,9 @@
 # frozen_string_literal: true
 
-require_relative '../core/utils/time'
-require_relative '../ruby_version'
-require_relative 'fatal_exceptions'
+require_relative "../core/utils/time"
+require_relative "../ruby_version"
+require_relative "fatal_exceptions"
+require_relative "capture_expression_evaluator"
 
 # rubocop:disable Lint/AssignmentInCondition
 # rubocop:disable Style/AndOr
@@ -84,6 +85,18 @@ module Datadog
       attr_reader :telemetry
       attr_reader :code_tracker
 
+      # The code tracker is a global singleton created lazily by
+      # DI.activate_tracking. When DI is enabled after boot via remote
+      # configuration this instrumenter was already built with a nil tracker;
+      # Component#start! assigns the now-current tracker here.
+      attr_writer :code_tracker
+
+      def capture_expression_evaluator
+        @capture_expression_evaluator ||= CaptureExpressionEvaluator.new(
+          settings: settings, serializer: serializer, logger: logger, telemetry: telemetry,
+        )
+      end
+
       # This is a substitute for Thread::Backtrace::Location
       # which does not have a public constructor.
       # Used for the fabricated stack frame for the method itself
@@ -179,7 +192,7 @@ module Datadog
           # invokes it with yield, which does not dispatch Proc#call, so a
           # user probe on Proc#call cannot intercept the trampoline, and no
           # Proc is allocated for the block.
-          if RubyVersion.is?('>= 3')
+          if RubyVersion.is?(">= 3")
             define_method(method_name) do |*args, **kwargs, &target_block| # steep:ignore NoMethod
               # steep:ignore FallbackAny below: Steep cannot narrow the
               # **kwargs parameter inside this define_method block, so it
@@ -520,8 +533,27 @@ module Datadog
             # they need to be serialized prior to method invocation.
             serialized_entry_args = if probe.capture_snapshot?
               serializer.serialize_args(args, kwargs, target_self,
-                depth: probe.max_capture_depth || settings.dynamic_instrumentation.max_capture_depth,
-                attribute_count: probe.max_capture_attribute_count || settings.dynamic_instrumentation.max_capture_attribute_count)
+                **probe.snapshot_serializer_limits(settings))
+            end
+
+            entry_capture_expressions = nil
+            entry_capture_evaluation_errors = nil
+            if probe.capture_entry_expressions?
+              begin
+                entry_context = Context.new(
+                  probe: probe, settings: settings, serializer: serializer,
+                  target_self: target_self,
+                  locals: serializer.combine_args(args, kwargs, target_self),
+                )
+                entry_capture_expressions, entry_capture_evaluation_errors =
+                  capture_expression_evaluator.evaluate(probe, entry_context)
+              rescue Exception => exc # standard:disable Lint/RescueException
+                Datadog::DI.reraise_if_fatal(exc)
+                raise if settings.dynamic_instrumentation.internal.propagate_all_exceptions
+
+                logger.debug { "di: error evaluating entry-time capture expressions: #{exc.class}: #{exc.message}" }
+                telemetry&.report(exc, description: "Error evaluating entry-time capture expressions")
+              end
             end
             # We intentionally do not use Core::Utils::Time.get_time
             # here because the time provider may be overridden by the
@@ -584,9 +616,14 @@ module Datadog
             caller_locs = method_frame + (caller_locations(2) || [])
             # TODO capture arguments at exit
 
-            context = Context.new(locals: nil, target_self: target_self,
+            capture_expression_locals = if probe.capture_expressions_only?
+              serializer.combine_args(args, kwargs, target_self)
+            end
+            context = Context.new(locals: capture_expression_locals, target_self: target_self,
               probe: probe, settings: settings, serializer: serializer,
               serialized_entry_args: serialized_entry_args,
+              entry_capture_expressions: entry_capture_expressions,
+              entry_capture_evaluation_errors: entry_capture_evaluation_errors,
               caller_locations: caller_locs,
               return_value: rv, duration: duration, exception: exc,)
 
@@ -663,7 +700,7 @@ module Datadog
       #
       # Defined only on Ruby < 3; the Ruby 3+ wrapper captures keyword
       # arguments directly and never calls this.
-      if RubyVersion.is?('< 3')
+      if RubyVersion.is?("< 3")
         def kwargs_from_splat(args)
           last = args.last
           if DI.hash?(last)
@@ -829,8 +866,8 @@ module Datadog
                 Utils.path_matches_suffix?(path, working_suffix, case_insensitive: case_insensitive)
               end
               break if found
-              break unless working_suffix.include?('/')
-              working_suffix.sub!(%r{.*/+}, '')
+              break unless working_suffix.include?("/")
+              working_suffix.sub!(%r{.*/+}, "")
             end
             break if found
           end
@@ -844,7 +881,11 @@ module Datadog
         has_per_method = code_tracker&.send(:instance_variable_defined?, :@per_method_registry) &&
           code_tracker.send(:per_method_registry).key?(loaded_path)
 
-        if has_per_method
+        if code_tracker.nil?
+          raise Error::DITargetNotInRegistry,
+            "File #{loaded_path} is loaded but code tracking is not active; " \
+            "line probes cannot be targeted."
+        elsif has_per_method
           raise Error::DITargetNotInRegistry,
             "File #{loaded_path} is loaded and has per-method iseqs, " \
             "but none cover line #{line_no}. " \

data/lib/datadog/di/logger.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require 'forwardable'
+require "forwardable"
 
 module Datadog
   module DI

data/lib/datadog/di/preload.rb

@@ -4,7 +4,7 @@
 # enable dynamic instrumentation for third-party libraries used by the
 # application.
 
-require_relative 'base'
+require_relative "base"
 
 # Code tracking is required for line probes to work; see the comments
 # on the activate_tracking methods in di.rb for further details.

data/lib/datadog/di/probe.rb

@@ -34,11 +34,17 @@ module Datadog
     class Probe
       KNOWN_TYPES = %i[log].freeze
 
+      EVALUATE_AT_VALUES = %i[entry exit].freeze
+
       def initialize(id:, type:,
         file: nil, line_no: nil, type_name: nil, method_name: nil,
         template: nil, template_segments: nil,
         capture_snapshot: false, max_capture_depth: nil,
-        max_capture_attribute_count: nil, condition: nil,
+        max_capture_attribute_count: nil,
+        max_capture_collection_size: nil, max_capture_string_length: nil,
+        capture_expressions: [],
+        evaluate_at: nil,
+        condition: nil,
         rate_limit: nil)
         # Perform some sanity checks here to detect unexpected attribute
         # combinations, in order to not do them in subsequent code.
@@ -81,9 +87,17 @@ module Datadog
         @capture_snapshot = !!capture_snapshot
         @max_capture_depth = max_capture_depth
         @max_capture_attribute_count = max_capture_attribute_count
+        @max_capture_collection_size = max_capture_collection_size
+        @max_capture_string_length = max_capture_string_length
+        @capture_expressions = capture_expressions || []
+        evaluate_at = :exit if evaluate_at.nil?
+        unless EVALUATE_AT_VALUES.include?(evaluate_at)
+          raise ArgumentError, "Unknown evaluate_at value: #{evaluate_at.inspect} (expected one of #{EVALUATE_AT_VALUES.inspect})"
+        end
+        @evaluate_at = evaluate_at
         @condition = condition
 
-        @rate_limit = rate_limit || (@capture_snapshot ? 1 : 5000)
+        @rate_limit = rate_limit || ((@capture_snapshot || !@capture_expressions.empty?) ? 1 : 5000)
         @rate_limiter = Datadog::Core::TokenBucket.new(@rate_limit)
 
         # At most one report per second.
@@ -120,6 +134,14 @@ module Datadog
       # the global default will be used.
       attr_reader :max_capture_attribute_count
 
+      attr_reader :max_capture_collection_size
+
+      attr_reader :max_capture_string_length
+
+      attr_reader :capture_expressions
+
+      attr_reader :evaluate_at
+
       # Rate limit in effect, in invocations per second. Always present.
       attr_reader :rate_limit
 
@@ -140,6 +162,29 @@ module Datadog
         @capture_snapshot
       end
 
+      def capture_expressions?
+        !@capture_expressions.empty?
+      end
+
+      def evaluate_at_entry?
+        evaluate_at == :entry
+      end
+
+      # Capture-expression mode: the probe captures expressions but not a full
+      # snapshot. Snapshot capture serializes its own values, so expression
+      # capture only runs when a snapshot is not being taken.
+      def capture_expressions_only?
+        capture_expressions? && !capture_snapshot?
+      end
+
+      def capture_entry_expressions?
+        capture_expressions_only? && evaluate_at_entry?
+      end
+
+      def snapshot_serializer_limits(settings)
+        CaptureLimits.resolve(expr_limits: nil, probe: self, settings: settings)
+      end
+
       # Returns whether the probe is a line probe.
       #
       # Method probes may still specify a file name (to aid in locating the

data/lib/datadog/di/probe_builder.rb

@@ -3,7 +3,9 @@
 # rubocop:disable Lint/AssignmentInCondition
 
 require_relative "probe"
-require_relative 'el'
+require_relative "capture_expression"
+require_relative "capture_limits"
+require_relative "el"
 
 module Datadog
   module DI
@@ -22,21 +24,36 @@ module Datadog
     module ProbeBuilder
       # Steep: https://github.com/soutaro/steep/issues/363
       PROBE_TYPES = { # steep:ignore IncompatibleAssignment
-        'LOG_PROBE' => :log,
+        "LOG_PROBE" => :log,
       }.freeze
 
       module_function
 
-      def build_from_remote_config(config)
+      CAPTURE_EXPRESSION_NAME_PATTERN = /\A[a-zA-Z0-9_?]+\z/
+
+      EVALUATE_AT_STRINGS = {
+        "ENTRY" => :entry,
+        "EXIT" => :exit,
+        "DEFAULT" => :exit,
+      }.freeze
+
+      def build_from_remote_config(config, logger:)
         # The validations here are not yet comprehensive.
-        type = config.fetch('type')
+        type = config.fetch("type")
         type_symbol = PROBE_TYPES[type] or raise ArgumentError, "Unrecognized probe type: #{type}"
-        cond = if cond_spec = config['when']
-          unless cond_spec['dsl'] && cond_spec['json']
+        cond = if cond_spec = config["when"]
+          unless cond_spec["dsl"] && cond_spec["json"]
             raise ArgumentError, "Malformed condition specification for probe: #{config}"
           end
-          compiled = EL::Compiler.new.compile(cond_spec['json'])
-          EL::Expression.new(cond_spec['dsl'], compiled)
+          compiled = EL::Compiler.new.compile(cond_spec["json"])
+          EL::Expression.new(cond_spec["dsl"], compiled)
+        end
+        capture_expressions = build_capture_expressions(config["captureExpressions"])
+        capture_expressions = dedup_capture_expressions(capture_expressions, config["id"], logger)
+        if !!config["captureSnapshot"] && !capture_expressions.empty?
+          logger.debug do
+            "di: probe #{config["id"]}: captureSnapshot=true wins over captureExpressions (n=#{capture_expressions.size})"
+          end
         end
         Probe.new(
           id: config.fetch("id"),
@@ -50,10 +67,14 @@ module Datadog
           # We should not be using the template for anything - we instead
           # use +segments+ - but keep the template for debugging.
           template: config["template"],
-          template_segments: build_template_segments(config['segments']),
+          template_segments: build_template_segments(config["segments"]),
           capture_snapshot: !!config["captureSnapshot"],
           max_capture_depth: config["capture"]&.[]("maxReferenceDepth"),
           max_capture_attribute_count: config["capture"]&.[]("maxFieldCount"),
+          max_capture_collection_size: config["capture"]&.[]("maxCollectionSize"),
+          max_capture_string_length: config["capture"]&.[]("maxLength"),
+          capture_expressions: capture_expressions,
+          evaluate_at: parse_evaluate_at(config["evaluateAt"], config["id"], logger),
           rate_limit: config["sampling"]&.[]("snapshotsPerSecond"),
           condition: cond,
         )
@@ -61,13 +82,74 @@ module Datadog
         raise ArgumentError, "Malformed remote configuration entry for probe: #{exc.class}: #{exc.message}: #{config}"
       end
 
+      def build_capture_expressions(raw)
+        return [] if raw.nil?
+        unless Array === raw
+          raise ArgumentError, "captureExpressions must be an array, got: #{raw.class}"
+        end
+        return [] if raw.empty?
+        raw.map do |entry|
+          unless Hash === entry
+            raise ArgumentError, "captureExpressions entry must be a hash, got: #{entry.class}"
+          end
+          name = entry["name"]
+          unless String === name && CAPTURE_EXPRESSION_NAME_PATTERN.match?(name)
+            raise ArgumentError, "captureExpressions entry name missing or invalid (must match #{CAPTURE_EXPRESSION_NAME_PATTERN.inspect}): #{name.inspect}"
+          end
+          expr_spec = entry["expr"]
+          unless Hash === expr_spec && expr_spec["dsl"] && expr_spec["json"]
+            raise ArgumentError, "captureExpressions entry #{name}: missing or malformed expr"
+          end
+          compiled = EL::Compiler.new.compile(expr_spec["json"])
+          expr = EL::Expression.new(expr_spec["dsl"], compiled)
+          limits = build_capture_limits(entry["capture"])
+          CaptureExpression.new(name: name, expr: expr, limits: limits)
+        end
+      end
+
+      def dedup_capture_expressions(capture_expressions, probe_id, logger)
+        return capture_expressions if capture_expressions.size < 2
+        by_name = {}
+        capture_expressions.each do |capture_expression|
+          by_name[capture_expression.name] = capture_expression
+        end
+        return capture_expressions if by_name.size == capture_expressions.size
+        logger.debug do
+          "di: probe #{probe_id}: collapsed duplicate captureExpressions names, kept last per name (#{capture_expressions.size} -> #{by_name.size})"
+        end
+        by_name.values
+      end
+
+      def build_capture_limits(raw)
+        return nil if raw.nil?
+        unless Hash === raw
+          raise ArgumentError, "capture-expression entry capture must be a hash, got: #{raw.class}"
+        end
+        CaptureLimits.new(
+          max_reference_depth: raw["maxReferenceDepth"],
+          max_collection_size: raw["maxCollectionSize"],
+          max_length: raw["maxLength"],
+          max_field_count: raw["maxFieldCount"],
+        )
+      end
+
+      def parse_evaluate_at(raw, probe_id, logger)
+        return :exit if raw.nil?
+        EVALUATE_AT_STRINGS[raw] || begin
+          logger.debug do
+            "di: probe #{probe_id}: unrecognized evaluateAt value #{raw.inspect}, defaulting to :exit"
+          end
+          :exit
+        end
+      end
+
       def build_template_segments(segments)
         segments&.map do |segment|
           if Hash === segment
-            if str = segment['str']
+            if str = segment["str"]
               str
-            elsif ast = segment['json']
-              unless dsl = segment['dsl']
+            elsif ast = segment["json"]
+              unless dsl = segment["dsl"]
                 raise ArgumentError, "Missing dsl for json in segment: #{segment}"
               end
               compiled = EL::Compiler.new.compile(ast)

data/lib/datadog/di/probe_file_loader/railtie.rb

@@ -6,7 +6,7 @@ module Datadog
       # Railtie class initializes dynamic instrumentation contrib code
       # in Rails environments.
       class Railtie < Rails::Railtie
-        initializer 'datadog.dynamic_instrumentation.load_probe_file' do |app|
+        initializer "datadog.dynamic_instrumentation.load_probe_file" do |app|
           ProbeFileLoader.load_now
         end
       end

data/lib/datadog/di/probe_file_loader.rb

@@ -1,16 +1,16 @@
 # frozen_string_literal: true
 
-require 'json'
+require "json"
 
-require_relative 'fatal_exceptions'
+require_relative "fatal_exceptions"
 
 module Datadog
   module DI
     module ProbeFileLoader
       module_function def load_now_or_later
         if Datadog::Core::Contrib::Rails::Utils.railtie_supported?
-          Datadog.logger.debug('di: loading probe_file_loader/railtie')
-          require_relative 'probe_file_loader/railtie'
+          Datadog.logger.debug("di: loading probe_file_loader/railtie")
+          require_relative "probe_file_loader/railtie"
         else
           load_now
         end
@@ -22,7 +22,7 @@ module Datadog
       module_function def load_now
         should_propagate = false
 
-        probe_file_path = DATADOG_ENV['DD_DYNAMIC_INSTRUMENTATION_PROBE_FILE']
+        probe_file_path = DATADOG_ENV["DD_DYNAMIC_INSTRUMENTATION_PROBE_FILE"]
         if probe_file_path.nil? || probe_file_path.empty?
           return
         end

data/lib/datadog/di/probe_manager.rb

@@ -2,7 +2,7 @@
 
 # rubocop:disable Lint/AssignmentInCondition
 
-require_relative 'fatal_exceptions'
+require_relative "fatal_exceptions"
 
 module Datadog
   module DI