Fix the tag key of ttfr to align with otlp semantic (#799)

Author: peachisai

apm-sniffer/apm-agent-core/src/main/java/org/apache/skywalking/apm/agent/core/context/tag/Tags.java

@@ -238,17 +238,17 @@ public static final class HTTP {
     /**
      * GEN_AI_STREAM_TTFR represents the time to first response (TTFR) for streaming operations.
      */
-    public static final StringTag GEN_AI_STREAM_TTFR = new StringTag(40, "gen_ai.stream.ttfr");
+    public static final StringTag GEN_AI_SERVER_TIME_TO_FIRST_TOKEN = new StringTag(40, "gen_ai.server.time_to_first_token");
 
     /**
      * GEN_AI_INPUT_MESSAGES represents the chat history provided to the model as an input.
      */
-    public static final StringTag GEN_AI_INPUT_MESSAGES = new StringTag(44, "gen_ai.input.messages");
+    public static final StringTag GEN_AI_INPUT_MESSAGES = new StringTag(41, "gen_ai.input.messages");
 
     /**
      * GEN_AI_OUTPUT_MESSAGES represents the messages returned by the model where each message represents a specific model response (choice, candidate).
      */
-    public static final StringTag GEN_AI_OUTPUT_MESSAGES = new StringTag(45, "gen_ai.output.messages");
+    public static final StringTag GEN_AI_OUTPUT_MESSAGES = new StringTag(42, "gen_ai.output.messages");
 
     /**
      * Creates a {@code StringTag} with the given key and cache it, if it's created before, simply return it without

apm-sniffer/apm-sdk-plugin/spring-plugins/spring-ai-1.x-plugin/src/main/java/org/apache/skywalking/apm/plugin/spring/ai/v1/ChatModelStreamInterceptor.java

@@ -150,7 +150,7 @@ private void recordTtfrIfFirstToken(AbstractSpan span, Generation generation, St
             return;
         }
         if (state.firstResponseReceived.compareAndSet(false, true)) {
-            Tags.GEN_AI_STREAM_TTFR.set(span, String.valueOf(System.currentTimeMillis() - state.startTime));
+            Tags.GEN_AI_SERVER_TIME_TO_FIRST_TOKEN.set(span, String.valueOf(System.currentTimeMillis() - state.startTime));
         }
     }
 

test/plugin/scenarios/spring-ai-1.x-scenario/config/expectedData.yaml

@@ -116,7 +116,7 @@ segmentItems:
               - { key: gen_ai.request.temperature, value: '0.7' }
               - { key: gen_ai.request.top_k, value: null }
               - { key: gen_ai.request.top_p, value: '0.9' }
-              - { key: gen_ai.stream.ttfr, value: not null }
+              - { key: gen_ai.server.time_to_first_token, value: not null }
               - { key: gen_ai.response.id, value: 'chatcmpl-fc1b64d3' }
               - { key: gen_ai.response.model, value: gpt-4.1-2025-04-14 }
               - { key: gen_ai.response.finish_reasons, value: STOP }