fix: Fix warning messages with healthcheck (#4793)

Signed-off-by: jthomson04 <[email protected]>

Author: jthomson04

components/src/dynamo/vllm/health_check.py

@@ -67,15 +67,14 @@ def __init__(self, engine_client=None):
         self.default_payload = {
             "token_ids": [bos_token_id],
             "sampling_options": {
-                "max_tokens": 1,
                 "temperature": 0.0,
             },
             "stop_conditions": {
+                "max_tokens": 1,
                 "stop": None,
                 "stop_token_ids": None,
                 "include_stop_str_in_output": False,
                 "ignore_eos": False,
-                "min_tokens": 0,
             },
         }
         super().__init__()

lib/runtime/src/health_check.rs

@@ -303,6 +303,11 @@ impl HealthCheckManager {
                             false
                         };
 
+                        tokio::spawn(async move {
+                            // We need to consume the rest of the stream to avoid warnings on the frontend.
+                            response_stream.for_each(|_| async {}).await;
+                        });
+
                         // Update health status based on response
                         system_health.lock().set_endpoint_health_status(
                             &endpoint_subject_owned,