Mark McLoughlinandOpenAI Codex
91fcec1235
Add `--logging-config.formatter=json` to enable the python-json-logger formatter without having to use a complete Python logging dictConfig. Signed-off-by: Mark McLoughlin <markmc@redhat.com> Co-authored-by: OpenAI Codex <noreply@openai.com>
110 lines
3.4 KiB
Python
110 lines
3.4 KiB
Python
# SPDX-License-Identifier: Apache-2.0
|
|
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
|
import asyncio
|
|
import json
|
|
from argparse import Namespace
|
|
from contextlib import asynccontextmanager
|
|
|
|
from fastapi import FastAPI
|
|
|
|
from vllm import envs
|
|
from vllm.config.logging import LogFormatter
|
|
from vllm.engine.protocol import EngineClient
|
|
from vllm.logger import init_logger
|
|
from vllm.utils.gc_utils import freeze_gc_heap
|
|
|
|
logger = init_logger(__name__)
|
|
|
|
|
|
def load_log_config(log_config_file: str | None) -> dict | None:
|
|
if not log_config_file:
|
|
return None
|
|
try:
|
|
with open(log_config_file) as f:
|
|
return json.load(f)
|
|
except Exception as e:
|
|
logger.warning(
|
|
"Failed to load log config from file %s: error %s", log_config_file, e
|
|
)
|
|
return None
|
|
|
|
|
|
def get_uvicorn_log_config(args: Namespace) -> dict | None:
|
|
"""Get the uvicorn log config based on the provided arguments.
|
|
|
|
Priority:
|
|
1. If log_config_file is specified, use it
|
|
2. If disable_access_log_for_endpoints or JSON formatting is specified,
|
|
create a config with the requested filter and format
|
|
3. Otherwise, return None (use uvicorn defaults)
|
|
"""
|
|
# First, try to load from file if specified
|
|
logging_config = getattr(args, "logging_config", None)
|
|
log_config_file = (
|
|
logging_config.pylogging_config_file
|
|
if logging_config is not None
|
|
else getattr(args, "log_config_file", None)
|
|
)
|
|
log_config = load_log_config(log_config_file)
|
|
if log_config is not None:
|
|
return log_config
|
|
|
|
formatter: LogFormatter = (
|
|
logging_config.formatter if logging_config is not None else "text"
|
|
)
|
|
if args.disable_access_log_for_endpoints or formatter == "json":
|
|
from vllm.logging_utils import create_uvicorn_log_config
|
|
|
|
excluded_paths = [
|
|
p.strip()
|
|
for p in (args.disable_access_log_for_endpoints or "").split(",")
|
|
if p.strip()
|
|
]
|
|
return create_uvicorn_log_config(
|
|
excluded_paths=excluded_paths,
|
|
log_level=args.uvicorn_log_level,
|
|
formatter=formatter,
|
|
)
|
|
|
|
return None
|
|
|
|
|
|
_running_tasks: set[asyncio.Task] = set()
|
|
|
|
|
|
@asynccontextmanager
|
|
async def lifespan(app: FastAPI):
|
|
try:
|
|
if app.state.log_stats:
|
|
engine_client: EngineClient = app.state.engine_client
|
|
|
|
async def _force_log():
|
|
while True:
|
|
await asyncio.sleep(envs.VLLM_LOG_STATS_INTERVAL)
|
|
await engine_client.do_log_stats()
|
|
|
|
task = asyncio.create_task(_force_log())
|
|
_running_tasks.add(task)
|
|
task.add_done_callback(_running_tasks.remove)
|
|
else:
|
|
task = None
|
|
|
|
# Mark the startup heap as static so that it's ignored by GC.
|
|
# Reduces pause times of oldest generation collections.
|
|
freeze_gc_heap()
|
|
try:
|
|
yield
|
|
finally:
|
|
if task is not None:
|
|
task.cancel()
|
|
for attr_name in (
|
|
"openai_serving_transcription",
|
|
"openai_serving_translation",
|
|
):
|
|
serving = getattr(app.state, attr_name, None)
|
|
if serving is not None and hasattr(serving, "shutdown"):
|
|
serving.shutdown()
|
|
finally:
|
|
# Ensure app state including engine ref is gc'd
|
|
del app.state
|