datadog 2.37.0 → 2.39.0

data/lib/datadog/di/probe_notification_builder.rb

@@ -2,7 +2,8 @@
 
 # rubocop:disable Lint/AssignmentInCondition
 
-require_relative 'fatal_exceptions'
+require_relative "fatal_exceptions"
+require_relative "capture_expression_evaluator"
 
 module Datadog
   module DI
@@ -10,43 +11,55 @@ module Datadog
     #
     # @api private
     class ProbeNotificationBuilder
-      def initialize(settings, serializer)
+      def initialize(settings, serializer, logger, telemetry: nil)
         @settings = settings
         @serializer = serializer
+        @logger = logger
+        @telemetry = telemetry
+        @capture_expression_evaluator = CaptureExpressionEvaluator.new(
+          settings: settings, serializer: serializer, logger: logger, telemetry: telemetry,
+        )
       end
 
       attr_reader :settings
+
       attr_reader :serializer
 
+      attr_reader :logger
+
+      attr_reader :telemetry
+
+      attr_reader :capture_expression_evaluator
+
       def build_received(probe)
         build_status(probe,
           message: "Probe #{probe.id} has been received correctly",
-          status: 'RECEIVED',)
+          status: "RECEIVED",)
       end
 
       def build_installed(probe)
         build_status(probe,
           message: "Probe #{probe.id} has been instrumented correctly",
-          status: 'INSTALLED',)
+          status: "INSTALLED",)
       end
 
       def build_emitting(probe)
         build_status(probe,
           message: "Probe #{probe.id} is emitting",
-          status: 'EMITTING',)
+          status: "EMITTING",)
       end
 
       def build_errored(probe, exc)
         build_status(probe,
           message: "Instrumentation for probe #{probe.id} failed: #{exc}",
-          status: 'ERROR',
+          status: "ERROR",
           exception: exc)
       end
 
       def build_disabled(probe, duration)
         build_status(probe,
           message: "Probe #{probe.id} was disabled because it consumed #{duration} seconds of CPU time in DI processing",
-          status: 'ERROR',)
+          status: "ERROR",)
       end
 
       # Duration is in seconds.
@@ -67,13 +80,13 @@ module Datadog
 
         # TODO also verify that non-capturing probe does not pass
         # snapshot or vars/args into this method
+        capture_expression_evaluation_errors = []
         captures = if probe.capture_snapshot?
+          snapshot_limits = probe.snapshot_serializer_limits(settings)
           if probe.method?
             return_arguments = {
-              "@return": serializer.serialize_value(context.return_value,
-                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),
-              self: serializer.serialize_value(context.target_self),
+              "@return": serializer.serialize_value(context.return_value, **snapshot_limits),
+              self: serializer.serialize_value(context.target_self, **snapshot_limits),
             }
             {
               entry: {
@@ -89,11 +102,38 @@ module Datadog
               lines: (locals = context.serialized_locals) && {
                 probe.line_no => {
                   locals: locals,
-                  arguments: {self: serializer.serialize_value(context.target_self)},
+                  arguments: {self: serializer.serialize_value(context.target_self, **snapshot_limits)},
                 },
               },
             }
           end
+        elsif probe.capture_expressions?
+          if probe.method?
+            if probe.evaluate_at_entry?
+              captured_block = context.entry_capture_expressions || {}
+              capture_expression_evaluation_errors = context.entry_capture_evaluation_errors || []
+              {
+                entry: {captureExpressions: captured_block},
+              }
+            else
+              captured_block, capture_expression_evaluation_errors =
+                capture_expression_evaluator.evaluate(probe, context)
+              {
+                return: {
+                  captureExpressions: captured_block,
+                  throwable: context.exception ? serialize_throwable(context.exception) : nil,
+                },
+              }
+            end
+          elsif probe.line?
+            captured_block, capture_expression_evaluation_errors =
+              capture_expression_evaluator.evaluate(probe, context)
+            {
+              lines: {
+                probe.line_no => {captureExpressions: captured_block},
+              },
+            }
+          end
         end
 
         message = nil
@@ -101,6 +141,7 @@ module Datadog
         if segments = probe.template_segments
           message, evaluation_errors = evaluate_template(segments, context)
         end
+        evaluation_errors.concat(capture_expression_evaluation_errors)
         build_snapshot_base(context,
           evaluation_errors: evaluation_errors, message: message,
           captures: captures)
@@ -135,9 +176,9 @@ module Datadog
         # appears to be completely ignored by the backend.
         # Note: The Go DI implementation does not send the top-level message
         # field at all when sending error statuses.
-        if status == 'ERROR'
+        if status == "ERROR"
           diagnostics[:exception] = { # steep:ignore
-            type: exception ? exception.class.name : 'Error',
+            type: exception ? exception.class.name : "Error",
             message: exception ? exception.message : message
           }
         end
@@ -146,7 +187,7 @@ module Datadog
           service: settings.service,
           timestamp: timestamp_now,
           message: message,
-          ddsource: 'dd_debugger',
+          ddsource: "dd_debugger",
           debugger: {
             diagnostics: diagnostics,
           },
@@ -189,7 +230,7 @@ module Datadog
           # Non-string constructor argument — return a redacted placeholder
           # rather than calling .to_s which could be customer code.
           # The exception class is already reported via the :type field.
-          '<REDACTED: not a string value>'
+          "<REDACTED: not a string value>"
         end
         # Prefer backtrace_locations (structured Location objects) over
         # backtrace (formatted strings that need regex parsing).
@@ -258,7 +299,7 @@ module Datadog
           if frame =~ BACKTRACE_FRAME_PATTERN
             {fileName: $1, function: $3, lineNumber: $2.to_i}
           else
-            {fileName: frame, function: '', lineNumber: 0}
+            {fileName: frame, function: "", lineNumber: 0}
           end
         end
       end
@@ -293,7 +334,7 @@ module Datadog
         payload = {
           service: settings.service,
           debugger: {
-            type: 'snapshot',
+            type: "snapshot",
             # Product can have three values: di, ld, er.
             # We do not currently implement exception replay.
             # There is currently no specification, and no consensus, for
@@ -305,7 +346,7 @@ module Datadog
             # except there is currently no consensus on said heuristics.
             # .NET always sends ld, other languages send nothing at the moment.
             # Don't send anything for the time being.
-            #product: 'di/ld',
+            # product: 'di/ld',
             snapshot: {
               id: SecureRandom.uuid,
               timestamp: timestamp,
@@ -315,7 +356,7 @@ module Datadog
                 version: 0,
                 location: location,
               },
-              language: 'ruby',
+              language: "ruby",
               # TODO add test coverage for callers being nil
               stack: stack,
               # System tests schema validation requires captures to
@@ -341,7 +382,7 @@ module Datadog
           # TODO add tests that the trace/span id is correctly propagated
           "dd.trace_id": active_trace&.id&.to_s,
           "dd.span_id": active_span&.id&.to_s,
-          ddsource: 'dd_debugger',
+          ddsource: "dd_debugger",
           message: message,
           timestamp: timestamp,
         }
@@ -374,7 +415,7 @@ module Datadog
             message: "#{exc.class}: #{exc.message}",
             expr: segment.dsl_expr, # steep:ignore NoMethod
           }
-          '[evaluation error]'
+          "[evaluation error]"
         end.join
         [message, evaluation_errors]
       end

data/lib/datadog/di/probe_notifier_worker.rb

@@ -1,7 +1,7 @@
 # frozen_string_literal: true
 
-require_relative '../core/semaphore'
-require_relative 'fatal_exceptions'
+require_relative "../core/semaphore"
+require_relative "fatal_exceptions"
 
 module Datadog
   module DI
@@ -231,7 +231,7 @@ module Datadog
         payload = probe_notification_builder.build_status(
           probe,
           message: "Probe #{probe.id} disabled: snapshot JSON encoding failed (#{exception.class}: #{exception})",
-          status: 'ERROR',
+          status: "ERROR",
           exception: exception,
         )
         add_status(payload, probe: probe)
@@ -241,13 +241,13 @@ module Datadog
         # DEV: The tags could be cached but they need to be recreated
         # when process forks (since the child receives new runtime IDs).
         Core::TagBuilder.tags(settings).merge(
-          'debugger_version' => Core::Environment::Identity.gem_datadog_version,
+          "debugger_version" => Core::Environment::Identity.gem_datadog_version,
         )
       end
 
       [
-        [:status, 'probe status'],
-        [:snapshot, 'snapshot'],
+        [:status, "probe status"],
+        [:snapshot, "snapshot"],
       ].each do |(event_type, event_name)|
         attr_reader "#{event_type}_queue"
 

data/lib/datadog/di/probe_repository.rb

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

data/lib/datadog/di/remote.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require_relative 'fatal_exceptions'
+require_relative "fatal_exceptions"
 
 module Datadog
   module DI
@@ -15,7 +15,7 @@ module Datadog
     # @api private
     module Remote
       class << self
-        PRODUCT = 'LIVE_DEBUGGING'
+        PRODUCT = "LIVE_DEBUGGING"
 
         # Declared here (not in Tracing::Remote::CAPABILITIES) so it is
         # registered only with the gated DI block in Capabilities#register:
@@ -228,7 +228,7 @@ module Datadog
           # we need to note it as being current so that we do not
           # try to remove instrumentation that is still supposed to be
           # active.
-          #current_probe_ids[probe_spec.fetch('id')] = true
+          # current_probe_ids[probe_spec.fetch('id')] = true
         rescue Exception => exc # standard:disable Lint/RescueException
           Datadog::DI.reraise_if_fatal(exc)
           raise if component.settings.dynamic_instrumentation.internal.propagate_all_exceptions
@@ -291,7 +291,7 @@ module Datadog
         def remove_probe(previous_content, component)
           # TODO test exception capture
           probe_spec = parse_content(previous_content)
-          probe_id = probe_spec.fetch('id')
+          probe_id = probe_spec.fetch("id")
           component.probe_manager.remove_probe(probe_id)
         rescue Exception => exc # standard:disable Lint/RescueException
           Datadog::DI.reraise_if_fatal(exc)

data/lib/datadog/di/serializer.rb

@@ -131,9 +131,12 @@ module Datadog
       # in upstream code.
       def serialize_args(args, kwargs, target_self,
         depth: settings.dynamic_instrumentation.max_capture_depth,
-        attribute_count: settings.dynamic_instrumentation.max_capture_attribute_count)
+        attribute_count: settings.dynamic_instrumentation.max_capture_attribute_count,
+        length: nil,
+        collection_size: nil)
         combined = combine_args(args, kwargs, target_self)
-        serialize_vars(combined, depth: depth, attribute_count: attribute_count)
+        serialize_vars(combined, depth: depth, attribute_count: attribute_count,
+          length: length, collection_size: collection_size)
       end
 
       # Serializes variables captured by a line probe.
@@ -142,9 +145,12 @@ module Datadog
       # of executed code.
       def serialize_vars(vars,
         depth: settings.dynamic_instrumentation.max_capture_depth,
-        attribute_count: settings.dynamic_instrumentation.max_capture_attribute_count)
+        attribute_count: settings.dynamic_instrumentation.max_capture_attribute_count,
+        length: nil,
+        collection_size: nil)
         vars.each_with_object({}) do |(k, v), agg|
-          agg[k] = serialize_value(v, name: k, depth: depth, attribute_count: attribute_count)
+          agg[k] = serialize_value(v, name: k, depth: depth, attribute_count: attribute_count,
+            length: length, collection_size: collection_size)
         end
       end
 
@@ -163,6 +169,8 @@ module Datadog
       def serialize_value(value, name: nil,
         depth: settings.dynamic_instrumentation.max_capture_depth,
         attribute_count: nil,
+        length: nil,
+        collection_size: nil,
         type: nil)
         attribute_count ||= settings.dynamic_instrumentation.max_capture_attribute_count
         cls = type || value.class
@@ -235,7 +243,8 @@ module Datadog
             #
             # Truncate binary data BEFORE escaping to avoid cutting mid-escape-sequence.
             # For regular strings, the limit is applied to string length in characters.
-            max = settings.dynamic_instrumentation.max_capture_string_length
+            length ||= settings.dynamic_instrumentation.max_capture_string_length
+            max = length
 
             if value.encoding == Encoding::BINARY || !value.valid_encoding?
               # Truncate binary data BEFORE escaping to avoid cutting mid-escape-sequence
@@ -263,7 +272,8 @@ module Datadog
             if depth <= 0
               serialized.update(notCapturedReason: "depth")
             else
-              max = settings.dynamic_instrumentation.max_capture_collection_size
+              collection_size ||= settings.dynamic_instrumentation.max_capture_collection_size
+              max = collection_size
               if max != 0 && value.length > max
                 serialized.update(notCapturedReason: "collectionSize", size: value.length)
                 # same steep failure with array slices.
@@ -271,7 +281,7 @@ module Datadog
                 value = value[0...max] || []
               end
               entries = value.map do |elt|
-                serialize_value(elt, depth: depth - 1)
+                serialize_value(elt, depth: depth - 1, length: length, collection_size: collection_size, attribute_count: attribute_count)
               end
               serialized.update(elements: entries)
             end
@@ -279,7 +289,8 @@ module Datadog
             if depth <= 0
               serialized.update(notCapturedReason: "depth")
             else
-              max = settings.dynamic_instrumentation.max_capture_collection_size
+              collection_size ||= settings.dynamic_instrumentation.max_capture_collection_size
+              max = collection_size
               cur = 0
               entries = []
               value.each do |k, v|
@@ -288,7 +299,8 @@ module Datadog
                   break
                 end
                 cur += 1
-                entries << [serialize_value(k, depth: depth - 1), serialize_value(v, name: k, depth: depth - 1)]
+                entries << [serialize_value(k, depth: depth - 1, length: length, collection_size: collection_size, attribute_count: attribute_count),
+                  serialize_value(v, name: k, depth: depth - 1, length: length, collection_size: collection_size, attribute_count: attribute_count)]
               end
               serialized.update(entries: entries)
             end
@@ -327,7 +339,7 @@ module Datadog
                   break
                 end
                 cur += 1
-                fields[ivar] = serialize_value(value.instance_variable_get(ivar), name: ivar, depth: depth - 1)
+                fields[ivar] = serialize_value(value.instance_variable_get(ivar), name: ivar, depth: depth - 1, length: length, collection_size: collection_size, attribute_count: attribute_count)
               end
               serialized.update(fields: fields)
             end
@@ -368,28 +380,28 @@ module Datadog
         # array allocations.
         case value
         when NilClass
-          'nil'
+          "nil"
         when Integer, Float, TrueClass, FalseClass, Time, Date
           value.to_s
         when String
           serialize_string_or_symbol_for_message(value)
         when Symbol
-          ':' + serialize_string_or_symbol_for_message(value) # steep:ignore ArgumentTypeMismatch
+          ":" + serialize_string_or_symbol_for_message(value) # steep:ignore ArgumentTypeMismatch
         when Array
-          return '...' if depth <= 0
+          return "..." if depth <= 0
 
           max = max_capture_collection_size_for_message
           if value.length > max
             value_ = value[0...max - 1] || []
-            value_ << '...'
+            value_ << "..."
             value_ << value[-1]
             value = value_
           end
-          '[' + value.map do |item|
+          "[" + value.map do |item|
             serialize_value_for_message(item, depth - 1)
-          end.join(', ') + ']'
+          end.join(", ") + "]"
         when Hash
-          return '...' if depth <= 0
+          return "..." if depth <= 0
 
           max = max_capture_collection_size_for_message
           keys = value.keys
@@ -405,11 +417,11 @@ module Datadog
           end
           if truncated
             serialized[serialized.length] = serialized[serialized.length - 1]
-            serialized[serialized.length - 2] = '...'
+            serialized[serialized.length - 2] = "..."
           end
           "{#{serialized.join(", ")}}"
         else
-          return '...' if depth <= 0
+          return "..." if depth <= 0
 
           vars = value.instance_variables
           truncated = false
@@ -427,10 +439,10 @@ module Datadog
           end
           if truncated
             serialized << serialized.last
-            serialized[-2] = '...'
+            serialized[-2] = "..."
           end
           serialized = if serialized.any?
-            ' ' + serialized.join(' ')
+            " " + serialized.join(" ")
           end
           "#<#{class_name(value.class)}#{serialized}>"
         end
@@ -489,7 +501,7 @@ module Datadog
             if max % 2 == 0
               upper += 1
             end
-            value[0...max / 2 - 1] + '...' + value[upper...length] # steep:ignore NoMethod
+            value[0...max / 2 - 1] + "..." + value[upper...length] # steep:ignore NoMethod
           end
         else
           value
@@ -531,7 +543,7 @@ module Datadog
           when 0x27 # '
             "\\'"
           when 0x5C # \
-            '\\\\'
+            "\\\\"
           when 0x20..0x7E # Printable ASCII (space through ~)
             byte.chr
           else

data/lib/datadog/di/transport/diagnostics.rb

@@ -1,11 +1,11 @@
 # frozen_string_literal: true
 
-require 'json'
+require "json"
 
-require_relative '../../core/transport/parcel'
-require_relative '../../core/transport/request'
-require_relative '../../core/transport/transport'
-require_relative 'http/diagnostics'
+require_relative "../../core/transport/parcel"
+require_relative "../../core/transport/request"
+require_relative "../../core/transport/transport"
+require_relative "http/diagnostics"
 
 module Datadog
   module DI

data/lib/datadog/di/transport/http/diagnostics.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require_relative '../../../core/transport/http/api/endpoint'
+require_relative "../../../core/transport/http/api/endpoint"
 
 module Datadog
   module DI
@@ -20,9 +20,9 @@ module Datadog
                 event_payload = Core::Vendor::Multipart::Post::UploadIO.new(
                   StringIO.new(env.request.parcel.data),
                   env.request.parcel.content_type,
-                  'event.json',
+                  "event.json",
                 )
-                env.form = {'event' => event_payload}
+                env.form = {"event" => event_payload}
 
                 super
               end

data/lib/datadog/di/transport/http/input.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require_relative '../../../core/transport/http/api/endpoint'
+require_relative "../../../core/transport/http/api/endpoint"
 
 module Datadog
   module DI
@@ -9,7 +9,7 @@ module Datadog
         module Input
           module API
             class Endpoint < Datadog::Core::Transport::HTTP::API::Endpoint
-              HEADER_CONTENT_TYPE = 'Content-Type'
+              HEADER_CONTENT_TYPE = "Content-Type"
 
               attr_reader \
                 :encoder

data/lib/datadog/di/transport/http.rb

@@ -1,9 +1,9 @@
 # frozen_string_literal: true
 
-require_relative '../../core/encoding'
-require_relative '../../core/transport/http'
-require_relative 'diagnostics'
-require_relative 'input'
+require_relative "../../core/encoding"
+require_relative "../../core/transport/http"
+require_relative "diagnostics"
+require_relative "input"
 
 module Datadog
   module DI
@@ -11,12 +11,12 @@ module Datadog
       # Namespace for HTTP transport components
       module HTTP
         DIAGNOSTICS = Diagnostics::API::Endpoint.new(
-          '/debugger/v1/diagnostics',
+          "/debugger/v1/diagnostics",
           Core::Encoding::JSONEncoder,
         )
 
         INPUT = Input::API::Endpoint.new(
-          '/debugger/v2/input',
+          "/debugger/v2/input",
           Core::Encoding::JSONEncoder,
         )
 
@@ -24,7 +24,7 @@ module Datadog
           # We used to use /debugger/v1/input, but now input
           # payloads should be going to the diagnostics endpoint
           # which I gather performs data redaction.
-          '/debugger/v1/diagnostics',
+          "/debugger/v1/diagnostics",
           Core::Encoding::JSONEncoder,
         )
 
@@ -40,7 +40,7 @@ module Datadog
             agent_settings: agent_settings,
             headers: headers,
           ) do |transport|
-            transport.api 'diagnostics', DIAGNOSTICS
+            transport.api "diagnostics", DIAGNOSTICS
 
             # Call block to apply any customization, if provided
             yield(transport) if block_given?
@@ -60,8 +60,8 @@ module Datadog
             agent_settings: agent_settings,
             headers: headers,
           ) do |transport|
-            transport.api 'input', INPUT, fallback: 'legacy_input', default: true
-            transport.api 'legacy_input', LEGACY_INPUT
+            transport.api "input", INPUT, fallback: "legacy_input", default: true
+            transport.api "legacy_input", LEGACY_INPUT
 
             # Call block to apply any customization, if provided
             yield(transport) if block_given?

data/lib/datadog/di/transport/input.rb

@@ -1,14 +1,14 @@
 # frozen_string_literal: true
 
-require_relative '../../core/chunker'
-require_relative '../../core/encoding'
-require_relative '../../core/tag_builder'
-require_relative '../../core/transport/parcel'
-require_relative '../../core/transport/request'
-require_relative '../../core/transport/transport'
-require_relative '../error'
-require_relative '../fatal_exceptions'
-require_relative 'http/input'
+require_relative "../../core/chunker"
+require_relative "../../core/encoding"
+require_relative "../../core/tag_builder"
+require_relative "../../core/transport/parcel"
+require_relative "../../core/transport/request"
+require_relative "../../core/transport/transport"
+require_relative "../error"
+require_relative "../fatal_exceptions"
+require_relative "http/input"
 
 module Datadog
   module DI

data/lib/datadog/di/utils.rb

@@ -86,7 +86,7 @@ module Datadog
       # forward slashes (DEBUG-5111). Used to normalize probe source paths
       # that originate from IDE tooling running on Windows.
       module_function def normalize_windows_separators(path)
-        path.tr('\\', '/')
+        path.tr("\\", "/")
       end
 
       # Returns whether the provided +path+ matches the user-designated
@@ -120,7 +120,7 @@ module Datadog
           suffix = suffix.downcase
         end
 
-        if suffix.start_with?('/')
+        if suffix.start_with?("/")
           path == suffix
         else
           # Exact match is not possible here, meaning any matching path
@@ -156,8 +156,8 @@ module Datadog
           return true if path_matches_suffix?(path, working_spec, case_insensitive: case_insensitive)
 
           loop do
-            break unless working_spec.include?('/')
-            working_spec.sub!(%r{.*/+}, '')
+            break unless working_spec.include?("/")
+            working_spec.sub!(%r{.*/+}, "")
             return true if path_matches_suffix?(path, working_spec, case_insensitive: case_insensitive)
           end
         end

data/lib/datadog/di.rb

@@ -1,8 +1,8 @@
 # frozen_string_literal: true
 
-require_relative 'di/configuration'
-require_relative 'di/extensions'
-require_relative 'di/remote'
+require_relative "di/configuration"
+require_relative "di/extensions"
+require_relative "di/remote"
 
 module Datadog
   # Namespace for Datadog dynamic instrumentation.
@@ -127,9 +127,9 @@ module Datadog
       #
       # @return [String, nil]
       private def unsupported_platform_reason
-        if RUBY_ENGINE != 'ruby'
+        if RUBY_ENGINE != "ruby"
           "MRI is required, but running on #{RUBY_ENGINE}"
-        elsif Datadog::RubyVersion.is?('< 2.6')
+        elsif Datadog::RubyVersion.is?("< 2.6")
           "Ruby 2.6+ is required, but running on #{RUBY_VERSION}"
         end
       end