Fix metrics and request tracing (TimeStats) (#11123)

python/pyproject.toml

@@ -14,18 +14,17 @@ classifiers = [
   "License :: OSI Approved :: Apache Software License",
 ]
 dependencies = [
-  "aiohttp",
-  "requests",
-  "tqdm",
-  "numpy",
   "IPython",
-  "setproctitle",
+  "aiohttp",
+  "anthropic>=0.20.0",
   "blobfile==3.0.0",
   "build",
   "compressed-tensors",
+  "cuda-python",
   "datasets",
   "einops",
   "fastapi",
+  "flashinfer_python==0.4.0rc3",
   "hf_transfer",
   "huggingface_hub",
   "interegular",
@@ -33,8 +32,10 @@ dependencies = [
   "modelscope",
   "msgspec",
   "ninja",
-  "openai==1.99.1",
+  "numpy",
+  "nvidia-cutlass-dsl==4.2.1",
   "openai-harmony==0.0.4",
+  "openai==1.99.1",
   "orjson",
   "outlines==0.1.11",
   "packaging",
@@ -42,32 +43,30 @@ dependencies = [
   "pillow",
   "prometheus-client>=0.20.0",
   "psutil",
+  "py-spy",
   "pybase64",
   "pydantic",
   "pynvml",
   "python-multipart",
   "pyzmq>=25.1.2",
+  "requests",
   "scipy",
   "sentencepiece",
+  "setproctitle",
+  "sgl-kernel==0.3.13",
   "soundfile==0.13.1",
-  "timm==1.0.16",
   "tiktoken",
+  "timm==1.0.16",
+  "torch==2.8.0",
+  "torch_memory_saver==0.0.8",
   "torchao==0.9.0",
+  "torchaudio==2.8.0",
+  "torchvision",
+  "tqdm",
   "transformers==4.56.1",
   "uvicorn",
   "uvloop",
-  "xgrammar==0.1.24",
-  "sgl-kernel==0.3.13",
-  "torch==2.8.0",
-  "torchaudio==2.8.0",
-  "torchvision",
-  "cuda-python",
-  "flashinfer_python==0.4.0rc3",
-  "openai==1.99.1",
-  "tiktoken",
-  "anthropic>=0.20.0",
-  "torch_memory_saver==0.0.8",
-  "nvidia-cutlass-dsl==4.2.1",
+  "xgrammar==0.1.24"
 ]
 
 [project.optional-dependencies]
@@ -79,15 +78,15 @@ test = [
   "matplotlib",
   "pandas",
   "peft",
-  "sentence_transformers",
   "pytest",
+  "sentence_transformers",
   "tabulate",
 ]
 tracing = [
-    "opentelemetry-sdk",
     "opentelemetry-api",
     "opentelemetry-exporter-otlp",
     "opentelemetry-exporter-otlp-proto-grpc",
+    "opentelemetry-sdk",
 ]
 all = ["sglang[test]", "sglang[decord]"]
 blackwell = ["sglang[test]", "sglang[decord]"]