Add soft watchdogs to debug soft hangs (#15023)

2025-12-13 10:41:35 +08:00
parent 487cf81a68
commit 313f59ad80
3 changed files with 23 additions and 6 deletions
--- a/python/sglang/srt/managers/scheduler.py
+++ b/python/sglang/srt/managers/scheduler.py
@@ -527,6 +527,8 @@ class Scheduler(
        self.watchdog = SchedulerWatchdog(
            self, watchdog_timeout=server_args.watchdog_timeout
        )
+        if (x := server_args.soft_watchdog_timeout) is not None:
+            self.soft_watchdog = SchedulerWatchdog(self, watchdog_timeout=x, soft=True)

        # Init memory saver, profiler and metric stats
        self.memory_saver_adapter = TorchMemorySaverAdapter.create(