Moved parsing to tracer settings plus other minor fixes

Author: kr-igor

tracer/src/Datadog.Trace/Configuration/TracerSettings.cs

@@ -14,6 +14,7 @@
 using Datadog.Trace.ClrProfiler.ServerlessInstrumentation;
 using Datadog.Trace.Configuration.ConfigurationSources.Telemetry;
 using Datadog.Trace.Configuration.Telemetry;
+using Datadog.Trace.DataStreamsMonitoring.TransactionTracking;
 using Datadog.Trace.LibDatadog;
 using Datadog.Trace.Logging;
 using Datadog.Trace.Logging.DirectSubmission;
@@ -587,7 +588,10 @@ not null when string.Equals(value, "otlp", StringComparison.OrdinalIgnoreCase) =
 
             DataStreamsTransactionExtractors = config
                                                   .WithKeys(ConfigurationKeys.DataStreamsMonitoring.TransactionExtractors)
-                                                  .AsString(string.Empty);
+                                                  .GetAs(
+                                                       defaultValue: new DefaultResult<IReadOnlyList<DataStreamsTransactionExtractor>>([], "[]"),
+                                                       converter: json => ParsingResult<IReadOnlyList<DataStreamsTransactionExtractor>>.Success(DataStreamsTransactionExtractor.ParseList(json)),
+                                                       validator: null);
 
             // no legacy headers if we are in "enbaled by default" state
             IsDataStreamsLegacyHeadersEnabled = config
@@ -1215,7 +1219,7 @@ not null when string.Equals(value, "otlp", StringComparison.OrdinalIgnoreCase) =
         /// <summary>
         /// Gets a raw value for DSM extractors
         /// </summary>
-        internal string DataStreamsTransactionExtractors { get; }
+        internal IReadOnlyList<DataStreamsTransactionExtractor> DataStreamsTransactionExtractors { get; }
 
         /// <summary>
         /// Gets a value indicating whether data streams schema extraction is enabled or not.

tracer/src/Datadog.Trace/DataStreamsMonitoring/TransactionTracking/DataStreamsExtractorRegistry.cs

@@ -4,43 +4,19 @@
 // </copyright>
 #nullable enable
 
-using System;
 using System.Collections.Generic;
-using Datadog.Trace.Logging;
 using Datadog.Trace.Util.Json;
 
 namespace Datadog.Trace.DataStreamsMonitoring.TransactionTracking;
 
 internal sealed class DataStreamsExtractorRegistry
 {
-    private static readonly IDatadogLogger Log = DatadogLogging.GetLoggerFor<DataStreamsExtractorRegistry>();
+    private readonly Dictionary<DataStreamsTransactionExtractor.ExtractorType, List<DataStreamsTransactionExtractor>> _extractors;
 
-    private readonly Dictionary<DataStreamsTransactionExtractor.ExtractorType, List<DataStreamsTransactionExtractor>> _extractors = new();
-
-    internal DataStreamsExtractorRegistry(string extractorsJson)
+    internal DataStreamsExtractorRegistry(IReadOnlyList<DataStreamsTransactionExtractor> extractors)
     {
-        if (string.IsNullOrWhiteSpace(extractorsJson))
-        {
-            return;
-        }
-
-        List<DataStreamsTransactionExtractor>? deserialized;
-        try
-        {
-            deserialized = JsonHelper.DeserializeObject<List<DataStreamsTransactionExtractor>>(extractorsJson);
-        }
-        catch (Exception ex)
-        {
-            Log.Warning(ex, "Failed to parse DD_DATA_STREAMS_TRANSACTION_EXTRACTORS value. Transaction tracking extractors will be disabled.");
-            return;
-        }
-
-        if (deserialized == null)
-        {
-            return;
-        }
-
-        foreach (var extractor in deserialized)
+        _extractors = new(extractors.Count);
+        foreach (var extractor in extractors)
         {
             var list = _extractors.GetValueOrDefault(extractor.ParsedType) ?? new();
             list.Add(extractor);

tracer/src/Datadog.Trace/DataStreamsMonitoring/TransactionTracking/DataStreamsTransactionExtractor.cs

@@ -5,6 +5,7 @@
 #nullable enable
 
 using System.Collections.Generic;
+using Datadog.Trace.Util.Json;
 using Datadog.Trace.Vendors.Newtonsoft.Json;
 
 namespace Datadog.Trace.DataStreamsMonitoring.TransactionTracking;
@@ -54,4 +55,18 @@ public ExtractorType ParsedType
             return _cachedType.Value;
         }
     }
+
+    public static IReadOnlyList<DataStreamsTransactionExtractor> ParseList(string json)
+    {
+        try
+        {
+            return JsonHelper.DeserializeObject<List<DataStreamsTransactionExtractor>>(json)
+                            ?.FindAll(e => e.ParsedType != ExtractorType.Unknown)
+                   ?? [];
+        }
+        catch
+        {
+            return [];
+        }
+    }
 }

tracer/src/Datadog.Trace/DataStreamsMonitoring/TransactionTracking/DataStreamsTransactionInfo.cs

@@ -58,9 +58,7 @@ internal DataStreamsTransactionInfo(byte[] idBytes, long timestamp, string check
         _checkpointId = Cache.GetOrAdd(checkpoint, _ => Interlocked.Increment(ref _counter));
     }
 
-    internal long TimestampNs { get => _timestamp; }
-
-    internal string TransactionId { get => Encoding.UTF8.GetString(_idBytes); }
+    internal long TimestampNs => _timestamp;
 
     private static byte[] Truncate(byte[] source)
     {

tracer/test/Datadog.Trace.Tests/DataStreamsMonitoring/DataStreamsExtractorRegistryTest.cs

@@ -12,29 +12,17 @@ namespace Datadog.Trace.Tests.DataStreamsMonitoring;
 
 public class DataStreamsExtractorRegistryTest
 {
-    [Theory]
-    [InlineData("not valid json")]
-    [InlineData("{\"not\": \"a list\"}")]
-    [InlineData("[{\"name\": \"n\", \"type\":")]  // truncated / incomplete
-    public void MalformedJson_DoesNotThrow_AndLeavesExtractorsEmpty(string json)
-    {
-        var registry = new DataStreamsExtractorRegistry(json);
-
-        registry.GetExtractorsByType(DataStreamsTransactionExtractor.ExtractorType.HttpOutHeaders).Should().BeNull();
-        registry.GetExtractorsByType(DataStreamsTransactionExtractor.ExtractorType.HttpInHeaders).Should().BeNull();
-    }
-
     [Fact]
     public void DeserializeCorrectly()
     {
-        var registry = new DataStreamsExtractorRegistry("[{\"name\": \"transaction-origin\", \"type\": \"HTTP_OUT_HEADERS\", \"value\": \"transaction-id\"}]");
+        var registry = FromJson("[{\"name\": \"transaction-origin\", \"type\": \"HTTP_OUT_HEADERS\", \"value\": \"transaction-id\"}]");
         registry.AsJson().Should().Be("{\"HttpOutHeaders\":[{\"name\":\"transaction-origin\",\"type\":\"HTTP_OUT_HEADERS\",\"value\":\"transaction-id\",\"ParsedType\":1}]}");
     }
 
     [Fact]
     public void GetExtractorsByType_ReturnsAllExtractors_ForSameType()
     {
-        var registry = new DataStreamsExtractorRegistry(
+        var registry = FromJson(
             "[" +
             "{\"name\": \"n1\", \"type\": \"HTTP_OUT_HEADERS\", \"value\": \"v1\"}," +
             "{\"name\": \"n2\", \"type\": \"HTTP_OUT_HEADERS\", \"value\": \"v2\"}" +
@@ -50,9 +38,11 @@ public void GetExtractorsByType_ReturnsAllExtractors_ForSameType()
     [Fact]
     public void GetExtractorsByType_DoesNotReturnExtractors_ForOtherType()
     {
-        var registry = new DataStreamsExtractorRegistry(
-            "[{\"name\": \"n1\", \"type\": \"HTTP_OUT_HEADERS\", \"value\": \"v1\"}]");
+        var registry = FromJson("[{\"name\": \"n1\", \"type\": \"HTTP_OUT_HEADERS\", \"value\": \"v1\"}]");
 
         registry.GetExtractorsByType(DataStreamsTransactionExtractor.ExtractorType.HttpInHeaders).Should().BeNull();
     }
+
+    private static DataStreamsExtractorRegistry FromJson(string json)
+        => new(DataStreamsTransactionExtractor.ParseList(json));
 }

tracer/test/Datadog.Trace.Tests/DataStreamsMonitoring/DataStreamsTransactionExtractorTests.cs

@@ -18,21 +18,30 @@ public class DataStreamsTransactionExtractorTests
     [InlineData("HTTP_IN_HEADERS",       2)]
     [InlineData("KAFKA_CONSUME_HEADERS", 3)]
     [InlineData("KAFKA_PRODUCE_HEADERS", 4)]
-    [InlineData("UNKNOWN_STUFF",         0)]
-    [InlineData("",                      0)]
-    public void ExtractorType_ReturnsCorrectType_ForTypeString(string stringType, int expectedInt)
+    public void ExtractorType_ReturnsCorrectType_ForKnownTypeString(string stringType, int expectedInt)
     {
         var expected = (DataStreamsTransactionExtractor.ExtractorType)expectedInt;
         var json = $"[{{\"name\": \"n\", \"type\": \"{stringType}\", \"value\": \"v\"}}]";
-        var registry = new DataStreamsExtractorRegistry(json);
+        var registry = new DataStreamsExtractorRegistry(DataStreamsTransactionExtractor.ParseList(json));
         registry.GetExtractorsByType(expected).Should().ContainSingle()
                 .Which.ParsedType.Should().Be(expected);
     }
 
+    [Theory]
+    [InlineData("UNKNOWN_STUFF")]
+    [InlineData("")]
+    public void ExtractorType_SkipsItem_ForUnknownTypeString(string stringType)
+    {
+        var json = $"[{{\"name\": \"n\", \"type\": \"{stringType}\", \"value\": \"v\"}}]";
+        var registry = new DataStreamsExtractorRegistry(DataStreamsTransactionExtractor.ParseList(json));
+        registry.GetExtractorsByType(DataStreamsTransactionExtractor.ExtractorType.Unknown).Should().BeNull();
+    }
+
     [Fact]
     public void ExtractorType_ReturnsSameValue_OnMultipleCalls()
     {
-        var registry = new DataStreamsExtractorRegistry("[{\"name\": \"n\", \"type\": \"HTTP_OUT_HEADERS\", \"value\": \"v\"}]");
+        var json = "[{\"name\": \"n\", \"type\": \"HTTP_OUT_HEADERS\", \"value\": \"v\"}]";
+        var registry = new DataStreamsExtractorRegistry(DataStreamsTransactionExtractor.ParseList(json));
         var extractor = registry.GetExtractorsByType(DataStreamsTransactionExtractor.ExtractorType.HttpOutHeaders)!.Single();
         extractor.ParsedType.Should().Be(extractor.ParsedType);
     }