[APIServer] support define gunicorn timeout (#4496)

* [BUGFIX] clear request #4286

* [BugFix] support define gunicorn timeout

* Update utils.py

* Update utils.py

---------

Co-authored-by: ltd0924 <luotingdan@baidu.com>

fastdeploy/entrypoints/openai/api_server.py

@@ -503,8 +503,8 @@ def launch_api_server() -> None:
         "workers": args.workers,
         "worker_class": "uvicorn.workers.UvicornWorker",
         "loglevel": "info",
-        "log_config": UVICORN_CONFIG,
-        "timeout_graceful_shutdown": args.timeout_graceful_shutdown,
+        "graceful_timeout": args.timeout_graceful_shutdown,
+        "timeout": args.timeout,
     }
 
     try:

fastdeploy/entrypoints/openai/utils.py

@@ -227,7 +227,14 @@ def make_arg_parser(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
         "--timeout-graceful-shutdown",
         default=0,
         type=int,
-        help="timeout for graceful shutdown in seconds (used by uvicorn)",
+        help="timeout for graceful shutdown in seconds (used by gunicorn).Setting it to 0 has the effect of infinite timeouts by disabling timeouts for all workers entirely.",
+    )
+
+    parser.add_argument(
+        "--timeout",
+        default=0,
+        type=int,
+        help="Workers silent for more than this many seconds are killed and restarted.Value is a positive number or 0. Setting it to 0 has the effect of infinite timeouts by disabling timeouts for all workers entirely.",
     )
 
     parser = EngineArgs.add_cli_args(parser)