From 89733111fa85edbc939fc9a43d6cb338c7fd0cea Mon Sep 17 00:00:00 2001
From: Li Wang <wangli858794774@gmail.com>
Date: Wed, 10 Dec 2025 09:24:19 +0800
Subject: [PATCH] [Nightly] Optimize nightly online test logger info (#4798)

### What this PR does / why we need it?
This patch do some tiny optimization for nightly ci:

1. Polling the frequency with which the service prints logs when it
starts up in order to obtain useful information more quickly.
2. Shorten the timeout for waiting server

- vLLM version: v0.12.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9

---------

Signed-off-by: wangli <wangli858794774@gmail.com>
---
 tests/e2e/conftest.py                         | 31 +++++++++++--------
 .../e2e/nightly/multi_node/test_multi_node.py |  2 +-
 2 files changed, 19 insertions(+), 14 deletions(-)

diff --git a/tests/e2e/conftest.py b/tests/e2e/conftest.py
index 8665aa91..9d0e709a 100644
--- a/tests/e2e/conftest.py
+++ b/tests/e2e/conftest.py
@@ -224,53 +224,58 @@ class RemoteOpenAIServer:
         # Then wait for all api_server nodes
         self._wait_for_multiple_servers(targets=targets, timeout=timeout)
 
-    def _wait_for_multiple_servers(self, targets, timeout: float):
+    def _wait_for_multiple_servers(self,
+                                   targets,
+                                   timeout: float,
+                                   log_interval: float = 30.0):
         """
         targets: List[(node_ip, url)]
+        log_interval
         """
         start = time.time()
         client = requests
 
-        # track readiness
         ready = {node_ip: False for node_ip, _ in targets}
 
-        # polling loop
+        last_log_time = 0.0
+
         while True:
+            now = time.time()
             all_ready = True
+            should_log = (now - last_log_time) >= log_interval
 
             for node_ip, url in targets:
                 if ready[node_ip]:
-                    continue  # already ready
+                    continue
 
                 try:
                     resp = client.get(url)
                     if resp.status_code == 200:
                         ready[node_ip] = True
                         logger.info(f"[READY] Node {node_ip} is ready.")
-                    else:
-                        all_ready = False
-                        logger.info(f"[WAIT] {url}: HTTP {resp.status_code}")
                 except RequestException:
                     all_ready = False
-                    logger.info(f"[WAIT] {url}: connection failed")
+                    if should_log:
+                        logger.info(f"[WAIT] {url}: connection failed")
 
-                    # underlying process died?
+                    # check unexpected exit
                     result = self._poll()
                     if result is not None and result != 0:
                         raise RuntimeError(
                             f"Server at {node_ip} exited unexpectedly."
                         ) from None
 
-            # if all nodes ready, exit
+            if should_log:
+                last_log_time = now
+
             if all_ready:
                 break
 
-            # check timeout
-            if time.time() - start > timeout:
+            if now - start > timeout:
                 not_ready_nodes = [n for n, ok in ready.items() if not ok]
                 self._terminate_server()
                 raise RuntimeError(
-                    f"Timeout: these nodes did not become ready: {not_ready_nodes}"
+                    f"Timeout: these nodes did not become ready: {not_ready_nodes} in time: {timeout}s"
                 ) from None
 
             time.sleep(5)
diff --git a/tests/e2e/nightly/multi_node/test_multi_node.py b/tests/e2e/nightly/multi_node/test_multi_node.py
index 66745a97..054ea31e 100644
--- a/tests/e2e/nightly/multi_node/test_multi_node.py
+++ b/tests/e2e/nightly/multi_node/test_multi_node.py
@@ -114,7 +114,7 @@ async def test_multi_node() -> None:
                 proxy_port=proxy_port,
                 disaggregated_prefill=disaggregated_prefill,
                 nodes_info=nodes_info,
-                max_wait_seconds=2000,
+                max_wait_seconds=1200,
         ) as remote_server:
             if config.is_master:
                 port = proxy_port if disaggregated_prefill else server_port