[Bugfix] Disable the statslogger if the api_server_count is greater than 1 (#22227)

Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
Co-authored-by: Nick Hill <nhill@redhat.com>

vllm/v1/engine/async_llm.py

@@ -143,6 +143,7 @@ class AsyncLLM(EngineClient):
                 engine_idxs=self.engine_core.engine_ranks_managed,
                 custom_stat_loggers=stat_loggers,
                 enable_default_loggers=log_stats,
+                client_count=client_count,
             )
             self.logger_manager.log_engine_initialized()
 

vllm/v1/metrics/loggers.py

@@ -652,6 +652,7 @@ class StatLoggerManager:
         engine_idxs: Optional[list[int]] = None,
         custom_stat_loggers: Optional[list[StatLoggerFactory]] = None,
         enable_default_loggers: bool = True,
+        client_count: int = 1,
     ):
         self.engine_idxs = engine_idxs if engine_idxs else [0]
 
@@ -660,6 +661,11 @@ class StatLoggerManager:
             factories.extend(custom_stat_loggers)
 
         if enable_default_loggers and logger.isEnabledFor(logging.INFO):
+            if client_count > 1:
+                logger.warning(
+                    "AsyncLLM created with api_server_count more than 1; "
+                    "disabling stats logging to avoid incomplete stats.")
+            else:
                 factories.append(LoggingStatLogger)
 
         # engine_idx: StatLogger