Generalize size limits based on envelope item type (#2421)

* Fix spec file location

* Extract Envelope::Item to its own file

* Add Envelope::Item#size_limit

* Make Envelope::Item#{headers,payload} just readers

We don't set headers and payloads after initializing an item, so these
accessors were not needed.

* Memoize Envelope::Item#type in the constructor

It's faster like that given that we rely on type value in multiple
places

* Memoize Envelope::Item#data_category

Same as with type, we rely on this value in multiple places so it makes
no sense to keep calculating it multiple times

* Add custom size limit for profile items

* Update CHANGELOG

CHANGELOG.md

@@ -1,3 +1,9 @@
+## Unreleased
+
+### Internal
+
+- Profile items have bigger size limit now ([#2421](https://github.com/getsentry/sentry-ruby/pull/2421))
+
 ## 5.20.1
 
 ### Bug Fixes

sentry-ruby/lib/sentry/envelope.rb

@@ -3,91 +3,6 @@
 module Sentry
   # @api private
   class Envelope
-    class Item
-      STACKTRACE_FRAME_LIMIT_ON_OVERSIZED_PAYLOAD = 500
-      MAX_SERIALIZED_PAYLOAD_SIZE = 1024 * 1000
-
-      attr_accessor :headers, :payload
-
-      def initialize(headers, payload)
-        @headers = headers
-        @payload = payload
-      end
-
-      def type
-        @headers[:type] || "event"
-      end
-
-      # rate limits and client reports use the data_category rather than envelope item type
-      def self.data_category(type)
-        case type
-        when "session", "attachment", "transaction", "profile", "span" then type
-        when "sessions" then "session"
-        when "check_in" then "monitor"
-        when "statsd", "metric_meta" then "metric_bucket"
-        when "event" then "error"
-        when "client_report" then "internal"
-        else "default"
-        end
-      end
-
-      def data_category
-        self.class.data_category(type)
-      end
-
-      def to_s
-        [JSON.generate(@headers), @payload.is_a?(String) ? @payload : JSON.generate(@payload)].join("\n")
-      end
-
-      def serialize
-        result = to_s
-
-        if result.bytesize > MAX_SERIALIZED_PAYLOAD_SIZE
-          remove_breadcrumbs!
-          result = to_s
-        end
-
-        if result.bytesize > MAX_SERIALIZED_PAYLOAD_SIZE
-          reduce_stacktrace!
-          result = to_s
-        end
-
-        [result, result.bytesize > MAX_SERIALIZED_PAYLOAD_SIZE]
-      end
-
-      def size_breakdown
-        payload.map do |key, value|
-          "#{key}: #{JSON.generate(value).bytesize}"
-        end.join(", ")
-      end
-
-      private
-
-      def remove_breadcrumbs!
-        if payload.key?(:breadcrumbs)
-          payload.delete(:breadcrumbs)
-        elsif payload.key?("breadcrumbs")
-          payload.delete("breadcrumbs")
-        end
-      end
-
-      def reduce_stacktrace!
-        if exceptions = payload.dig(:exception, :values) || payload.dig("exception", "values")
-          exceptions.each do |exception|
-            # in most cases there is only one exception (2 or 3 when have multiple causes), so we won't loop through this double condition much
-            traces = exception.dig(:stacktrace, :frames) || exception.dig("stacktrace", "frames")
-
-            if traces && traces.size > STACKTRACE_FRAME_LIMIT_ON_OVERSIZED_PAYLOAD
-              size_on_both_ends = STACKTRACE_FRAME_LIMIT_ON_OVERSIZED_PAYLOAD / 2
-              traces.replace(
-                traces[0..(size_on_both_ends - 1)] + traces[-size_on_both_ends..-1],
-              )
-            end
-          end
-        end
-      end
-    end
-
     attr_accessor :headers, :items
 
     def initialize(headers = {})
@@ -108,3 +23,5 @@ def event_id
     end
   end
 end
+
+require_relative "envelope/item"

sentry-ruby/lib/sentry/envelope/item.rb

@@ -0,0 +1,88 @@
+# frozen_string_literal: true
+
+module Sentry
+  # @api private
+  class Envelope::Item
+    STACKTRACE_FRAME_LIMIT_ON_OVERSIZED_PAYLOAD = 500
+    MAX_SERIALIZED_PAYLOAD_SIZE = 1024 * 1000
+
+    SIZE_LIMITS = Hash.new(MAX_SERIALIZED_PAYLOAD_SIZE).update(
+      "profile" => 1024 * 1000 * 50
+    )
+
+    attr_reader :size_limit, :headers, :payload, :type, :data_category
+
+    # rate limits and client reports use the data_category rather than envelope item type
+    def self.data_category(type)
+      case type
+      when "session", "attachment", "transaction", "profile", "span" then type
+      when "sessions" then "session"
+      when "check_in" then "monitor"
+      when "statsd", "metric_meta" then "metric_bucket"
+      when "event" then "error"
+      when "client_report" then "internal"
+      else "default"
+      end
+    end
+
+    def initialize(headers, payload)
+      @headers = headers
+      @payload = payload
+      @type = headers[:type] || "event"
+      @data_category = self.class.data_category(type)
+      @size_limit = SIZE_LIMITS[type]
+    end
+
+    def to_s
+      [JSON.generate(@headers), @payload.is_a?(String) ? @payload : JSON.generate(@payload)].join("\n")
+    end
+
+    def serialize
+      result = to_s
+
+      if result.bytesize > size_limit
+        remove_breadcrumbs!
+        result = to_s
+      end
+
+      if result.bytesize > size_limit
+        reduce_stacktrace!
+        result = to_s
+      end
+
+      [result, result.bytesize > size_limit]
+    end
+
+    def size_breakdown
+      payload.map do |key, value|
+        "#{key}: #{JSON.generate(value).bytesize}"
+      end.join(", ")
+    end
+
+    private
+
+    def remove_breadcrumbs!
+      if payload.key?(:breadcrumbs)
+        payload.delete(:breadcrumbs)
+      elsif payload.key?("breadcrumbs")
+        payload.delete("breadcrumbs")
+      end
+    end
+
+    def reduce_stacktrace!
+      if exceptions = payload.dig(:exception, :values) || payload.dig("exception", "values")
+        exceptions.each do |exception|
+          # in most cases there is only one exception (2 or 3 when have multiple causes), so we won't loop through this double condition much
+          traces = exception.dig(:stacktrace, :frames) || exception.dig("stacktrace", "frames")
+
+          if traces && traces.size > STACKTRACE_FRAME_LIMIT_ON_OVERSIZED_PAYLOAD
+            size_on_both_ends = STACKTRACE_FRAME_LIMIT_ON_OVERSIZED_PAYLOAD / 2
+            traces.replace(
+              traces[0..(size_on_both_ends - 1)] + traces[-size_on_both_ends..-1],
+            )
+          end
+        end
+      end
+    end
+  end
+end

sentry-ruby/spec/sentry/transport_spec.rb

@@ -145,6 +145,37 @@
           expect(profile_payload).to eq(profile.to_json)
         end
       end
+
+      context "allows bigger item size" do
+        let(:profile) do
+          {
+            environment: "test",
+            release: "release",
+            profile: {
+              frames: Array.new(10000) { |i| { function: "function_#{i}", filename: "file_#{i}", lineno: i } },
+              stacks: Array.new(10000) { |i| [i] },
+              samples: Array.new(10000) { |i| { stack_id: i, elapsed_since_start_ns: i * 1000, thread_id: i % 10 } }
+            }
+          }
+        end
+
+        let(:event_with_profile) do
+          event.profile = profile
+          event
+        end
+
+        let(:envelope) { subject.envelope_from_event(event_with_profile) }
+
+        it "adds profile item to envelope" do
+          result, _ = subject.serialize_envelope(envelope)
+
+          _profile_header, profile_payload_json = result.split("\n").last(2)
+
+          profile_payload = JSON.parse(profile_payload_json)
+
+          expect(profile_payload["profile"]).to_not be(nil)
+        end
+      end
     end
 
     context "client report" do