[Feature][TTS] Add Speech API streaming metrics (#6853)

Signed-off-by: XIN GAO <1037396230@qq.com>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>

tests/entrypoints/openai_api/test_serving_speech.py

@@ -3287,7 +3287,7 @@ class TestStreamingResponse:
         finalized_tts_params = {"_qwen3_tts_effective_max_tokens": [192]}
         captured: dict = {}
 
-        async def prepare(_request, request_id=None):
+        async def prepare(_request, request_id=None, arrival_time=None):
             return request_id, object(), finalized_tts_params
 
         async def generate_chunks(
@@ -3296,6 +3296,7 @@ class TestStreamingResponse:
             _response_format="pcm",
             raw_request=None,
             request_start_s=None,
+            request_arrival_ts=None,
             include_sample_rate=False,
             usage_acc=None,
             tts_params=None,