Collect process tags for tracing

Author: amarziali

dd-java-agent/instrumentation/graal/native-image/src/main/java/datadog/trace/instrumentation/graal/nativeimage/NativeImageGeneratorRunnerInstrumentation.java

@@ -76,6 +76,7 @@ public static void onEnter(@Advice.Argument(value = 0, readOnly = false) String[
               + "datadog.trace.api.Platform:rerun,"
               + "datadog.trace.api.Platform$Captured:build_time,"
               + "datadog.trace.api.env.CapturedEnvironment:build_time,"
+              + "datadog.trace.api.env.CapturedEnvironment$ProcessInfo:build_time,"
               + "datadog.trace.api.ConfigCollector:rerun,"
               + "datadog.trace.api.ConfigDefaults:build_time,"
               + "datadog.trace.api.ConfigOrigin:build_time,"

dd-java-agent/instrumentation/spring-boot/src/main/java/datadog/trace/instrumentation/springboot/SpringApplicationInstrumentation.java

@@ -7,7 +7,9 @@
 import datadog.trace.agent.tooling.Instrumenter;
 import datadog.trace.agent.tooling.InstrumenterModule;
 import datadog.trace.api.Config;
+import datadog.trace.api.ProcessTags;
 import datadog.trace.bootstrap.instrumentation.api.AgentTracer;
+import java.util.Arrays;
 import net.bytebuddy.asm.Advice;
 import org.springframework.core.env.ConfigurableEnvironment;
 
@@ -60,6 +62,17 @@ public static void afterEnvironmentPostProcessed(
       final String applicationName = environment.getProperty("spring.application.name");
       if (applicationName != null && !applicationName.isEmpty()) {
         AgentTracer.get().updatePreferredServiceName(applicationName);
+        ProcessTags.addTag("springboot.application", applicationName);
+      }
+      final String[] profiles = environment.getActiveProfiles();
+      System.err.println(Arrays.toString(profiles));
+      if (profiles != null && profiles.length > 0) {
+        ProcessTags.addTag("springboot.profile", profiles[0]);
+      } else {
+        final String[] defaultProfiles = environment.getDefaultProfiles();
+        if (defaultProfiles != null && defaultProfiles.length > 0) {
+          ProcessTags.addTag("springboot.profile", defaultProfiles[0]);
+        }
       }
     }
   }

dd-java-agent/instrumentation/spring-boot/src/test/groovy/SpringBootApplicationTest.groovy

@@ -1,3 +1,5 @@
+import datadog.trace.api.ProcessTags
+
 import static datadog.trace.agent.test.utils.TraceUtils.runUnderTrace
 
 import datadog.trace.agent.test.AgentTestRunner
@@ -10,6 +12,7 @@ class SpringBootApplicationTest extends AgentTestRunner {
   protected void configurePreAgent() {
     super.configurePreAgent()
   }
+
   static class BeanWhoTraces implements InitializingBean {
 
     @Override
@@ -32,15 +35,19 @@ class SpringBootApplicationTest extends AgentTestRunner {
         }
       }
     })
+    and:
+    def processTags = ProcessTags.getTagsForSerialization()
+    assert processTags.toString() =~ ".+,springboot.application:$expectedService,springboot.profile:$expectedProfile"
     context != null
     cleanup:
     context?.stop()
     where:
-    expectedService | args
-    "application-name-from-args"          | new String[]{
-      "--spring.application.name=application-name-from-args"
+    expectedService                    | expectedProfile | args
+    "application-name-from-args"       | "prod"          | new String[]{
+      "--spring.application.name=application-name-from-args",
+      "--spring.profiles.active=prod,common",
     }
-    "application-name-from-properties"         | new String[0] // will load from properties
+    "application-name-from-properties" | "default"       | new String[0] // will load from properties
   }
 }
 

dd-java-agent/testing/src/main/groovy/datadog/trace/agent/test/AgentTestRunner.groovy

@@ -27,6 +27,7 @@ import datadog.trace.agent.tooling.bytebuddy.matcher.GlobalIgnores
 import datadog.trace.api.Config
 import datadog.trace.api.DDSpanId
 import datadog.trace.api.IdGenerationStrategy
+import datadog.trace.api.ProcessTags
 import datadog.trace.api.StatsDClient
 import datadog.trace.api.TraceConfig
 import datadog.trace.api.WellKnownTags
@@ -509,6 +510,7 @@ abstract class AgentTestRunner extends DDSpecification implements AgentBuilder.L
       ActiveSubsystems.APPSEC_ACTIVE = true
     }
     InstrumentationErrors.resetErrorCount()
+    ProcessTags.reset()
   }
 
   @Override

dd-trace-api/src/main/java/datadog/trace/api/ConfigDefaults.java

@@ -241,7 +241,8 @@ public final class ConfigDefaults {
   static final int DEFAULT_TELEMETRY_DEPENDENCY_RESOLUTION_QUEUE_SIZE = 100000;
 
   static final Set<String> DEFAULT_TRACE_EXPERIMENTAL_FEATURES_ENABLED =
-      new HashSet<>(asList("DD_TAGS", "DD_LOGS_INJECTION"));
+      new HashSet<>(
+          asList("DD_TAGS", "DD_LOGS_INJECTION", "EXPERIMENTAL_COLLECT_PROCESS_TAGS_ENABLED"));
 
   static final boolean DEFAULT_TRACE_128_BIT_TRACEID_GENERATION_ENABLED = true;
   static final boolean DEFAULT_TRACE_128_BIT_TRACEID_LOGGING_ENABLED = true;

dd-trace-api/src/main/java/datadog/trace/api/DDTags.java

@@ -98,4 +98,5 @@ public class DDTags {
   public static final String DECISION_MAKER_INHERITED = "_dd.dm.inherited";
   public static final String DECISION_MAKER_SERVICE = "_dd.dm.service";
   public static final String DECISION_MAKER_RESOURCE = "_dd.dm.resource";
+  public static final String PROCESS_TAGS = "_dd.process";
 }

dd-trace-api/src/main/java/datadog/trace/api/config/GeneralConfig.java

@@ -29,6 +29,9 @@ public final class GeneralConfig {
   @Deprecated // Use dd.tags instead
   public static final String GLOBAL_TAGS = "trace.global.tags";
 
+  public static final String EXPERIMENTAL_COLLECT_PROCESS_TAGS_ENABLED =
+      "experimental.collect.tags.enabled";
+
   public static final String LOG_LEVEL = "log.level";
   public static final String TRACE_DEBUG = "trace.debug";
   public static final String TRACE_TRIAGE = "trace.triage";

dd-trace-core/src/main/java/datadog/trace/core/tagprocessor/ProcessTagsAdder.java

@@ -0,0 +1,21 @@
+package datadog.trace.core.tagprocessor;
+
+import static datadog.trace.api.DDTags.PROCESS_TAGS;
+
+import datadog.trace.api.ProcessTags;
+import datadog.trace.bootstrap.instrumentation.api.AgentSpanLink;
+import datadog.trace.core.DDSpanContext;
+import java.util.List;
+import java.util.Map;
+
+public class ProcessTagsAdder implements TagsPostProcessor {
+  @Override
+  public Map<String, Object> processTags(
+      Map<String, Object> unsafeTags, DDSpanContext spanContext, List<AgentSpanLink> spanLinks) {
+    final CharSequence processTags = ProcessTags.getTagsForSerialization();
+    if (processTags != null) {
+      unsafeTags.put(PROCESS_TAGS, processTags);
+    }
+    return unsafeTags;
+  }
+}

dd-trace-core/src/main/java/datadog/trace/core/tagprocessor/TagsPostProcessorFactory.java

@@ -10,7 +10,7 @@ public final class TagsPostProcessorFactory {
 
   private static class Lazy {
     private static TagsPostProcessor create() {
-      final List<TagsPostProcessor> processors = new ArrayList<>(4);
+      final List<TagsPostProcessor> processors = new ArrayList<>(7);
       processors.add(new PeerServiceCalculator());
       if (addBaseService) {
         processors.add(new BaseServiceAdder(Config.get().getServiceName()));
@@ -32,6 +32,9 @@ private static TagsPostProcessor create() {
       if (Config.get().isAddSpanPointers("aws")) {
         processors.add(new SpanPointersProcessor());
       }
+      if (Config.get().isExperimentalCollectProcessTagsEnabled()) {
+        processors.add(new ProcessTagsAdder());
+      }
       return new PostProcessorChain(
           processors.toArray(processors.toArray(new TagsPostProcessor[0])));
     }
@@ -52,6 +55,7 @@ public static void withAddBaseService(boolean enabled) {
     addBaseService = enabled;
     Lazy.instance = Lazy.create();
   }
+
   /**
    * Mostly used for test purposes.
    *

dd-trace-core/src/test/groovy/datadog/trace/core/tagprocessor/ProcessTagsAdderTest.groovy

@@ -0,0 +1,27 @@
+package datadog.trace.core.tagprocessor
+
+import datadog.trace.api.ProcessTags
+import datadog.trace.core.DDSpanContext
+import datadog.trace.test.util.DDSpecification
+
+class ProcessTagsAdderTest extends DDSpecification {
+  def "should add _dd.process"() {
+    setup:
+    def adder = new ProcessTagsAdder()
+    def spanContext = Mock(DDSpanContext)
+
+    when:
+    def enrichedTags = adder.processTags([:], spanContext, [])
+
+    then:
+    def firstValue = enrichedTags.get("_dd.process")
+    assert !firstValue.toString().isEmpty()
+
+    when:
+    ProcessTags.addTag("test", "value")
+    enrichedTags = adder.processTags([:], spanContext, [])
+
+    then:
+    assert enrichedTags.get("_dd.process").toString() == "$firstValue,test:value"
+  }
+}

internal-api/src/main/java/datadog/trace/api/Config.java

@@ -150,6 +150,7 @@ public static String getHostName() {
   private final boolean peerServiceDefaultsEnabled;
   private final Map<String, String> peerServiceComponentOverrides;
   private final boolean removeIntegrationServiceNamesEnabled;
+  private final boolean experimentalCollectProcessTagsEnabled;
   private final Map<String, String> peerServiceMapping;
   private final Map<String, String> serviceMapping;
   private final Map<String, String> tags;
@@ -846,6 +847,8 @@ private Config(final ConfigProvider configProvider, final InstrumenterConfig ins
     // feature flag to remove fake services in v0
     removeIntegrationServiceNamesEnabled =
         configProvider.getBoolean(TRACE_REMOVE_INTEGRATION_SERVICE_NAMES_ENABLED, false);
+    experimentalCollectProcessTagsEnabled =
+        configProvider.getBoolean(EXPERIMENTAL_COLLECT_PROCESS_TAGS_ENABLED, false);
 
     peerServiceMapping = configProvider.getMergedMap(TRACE_PEER_SERVICE_MAPPING);
 
@@ -2091,6 +2094,10 @@ public Set<String> getExperimentalFeaturesEnabled() {
     return experimentalFeaturesEnabled;
   }
 
+  public boolean isExperimentalCollectProcessTagsEnabled() {
+    return experimentalCollectProcessTagsEnabled;
+  }
+
   public boolean isTraceEnabled() {
     return instrumenterConfig.isTraceEnabled();
   }

internal-api/src/main/java/datadog/trace/api/ProcessTags.java

@@ -0,0 +1,120 @@
+package datadog.trace.api;
+
+import datadog.trace.api.env.CapturedEnvironment;
+import datadog.trace.bootstrap.instrumentation.api.UTF8BytesString;
+import datadog.trace.util.TraceUtils;
+import java.nio.file.Path;
+import java.nio.file.Paths;
+import java.util.LinkedHashMap;
+import java.util.Map;
+import java.util.stream.Collectors;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+public class ProcessTags {
+  private static final Logger LOGGER = LoggerFactory.getLogger(ProcessTags.class);
+
+  private static class Lazy {
+    static final Map<String, String> TAGS = loadTags();
+    static volatile UTF8BytesString serializedForm;
+
+    private static Map<String, String> loadTags() {
+      Map<String, String> tags = new LinkedHashMap<>();
+      try {
+        fillBaseTags(tags);
+        fillJbossTags(tags);
+      } catch (Throwable t) {
+        LOGGER.debug("Unable to calculate default process tags", t);
+      }
+      return tags;
+    }
+
+    private static void insertSysPropIfPresent(
+        Map<String, String> tags, String propKey, String tagKey) {
+      String value = System.getProperty(propKey);
+      if (value != null) {
+        tags.put(tagKey, value);
+      }
+    }
+
+    private static boolean insertLastPathSegmentIfPresent(
+        Map<String, String> tags, String path, String tagKey) {
+      if (path == null || path.isEmpty()) {
+        return false;
+      }
+      try {
+        final Path p = Paths.get(path).getFileName();
+        if (p != null) {
+          tags.put(tagKey, p.toString());
+          return true;
+        }
+      } catch (Throwable ignored) {
+      }
+      return false;
+    }
+
+    private static void fillBaseTags(Map<String, String> tags) {
+      final CapturedEnvironment.ProcessInfo processInfo =
+          CapturedEnvironment.get().getProcessInfo();
+      if (processInfo.mainClass != null) {
+        tags.put("entrypoint.name", processInfo.mainClass);
+      }
+      if (processInfo.jarFile != null) {
+        final String jarName = processInfo.jarFile.getName();
+        tags.put("entrypoint.name", jarName.substring(0, jarName.length() - 4)); // strip .jar
+        insertLastPathSegmentIfPresent(tags, processInfo.jarFile.getParent(), "entrypoint.basedir");
+      }
+
+      insertLastPathSegmentIfPresent(tags, System.getProperty("user.dir"), "entrypoint.workdir");
+    }
+
+    private static void fillJbossTags(Map<String, String> tags) {
+      if (insertLastPathSegmentIfPresent(
+          tags, System.getProperty("jboss.home.dir"), "jboss.home")) {
+        insertSysPropIfPresent(tags, "jboss.server.name", "server.name");
+        tags.put(
+            "jboss.mode",
+            System.getProperties().containsKey("[Standalone]") ? "standalone" : "domain");
+      }
+    }
+
+    static synchronized UTF8BytesString calculateSerializedForm() {
+      if (serializedForm == null && !TAGS.isEmpty()) {
+        serializedForm =
+            UTF8BytesString.create(
+                TAGS.entrySet().stream()
+                    .map(entry -> entry.getKey() + ":" + TraceUtils.normalizeTag(entry.getValue()))
+                    .collect(Collectors.joining(",")));
+      }
+      return serializedForm;
+    }
+  }
+
+  private ProcessTags() {}
+
+  // need to be synchronized on writing. As optimization, it does not need to be sync on read.
+  public static synchronized void addTag(String key, String value) {
+    Lazy.TAGS.put(key, value);
+    Lazy.serializedForm = null;
+  }
+
+  public static UTF8BytesString getTagsForSerialization() {
+    final UTF8BytesString serializedForm = Lazy.serializedForm;
+    if (serializedForm != null) {
+      return serializedForm;
+    }
+    return Lazy.calculateSerializedForm();
+  }
+
+  /** Visible for testing. */
+  static void empty() {
+    Lazy.TAGS.clear();
+    Lazy.serializedForm = null;
+  }
+
+  /** Visible for testing. */
+  static void reset() {
+    empty();
+    Lazy.TAGS.putAll(Lazy.loadTags());
+  }
+}