add background health check to miles native router

miles/router/router.py

@@ -1,5 +1,7 @@
 import argparse
+import asyncio
 import json
+import logging
 
 import httpx
 import uvicorn
@@ -9,6 +11,8 @@
 
 from miles.utils.misc import load_function
 
+logger = logging.getLogger(__name__)
+
 
 def run_router(args):
     """
@@ -28,6 +32,7 @@ def __init__(self, args, verbose=False):
         self.verbose = verbose
 
         self.app = FastAPI()
+        self.app.add_event_handler("startup", self._start_background_health_check)
 
         # Worker information
         self.worker_urls: dict[str, int] = {}
@@ -63,9 +68,39 @@ def _setup_routes(self):
         # Catch-all route for proxying to SGLang - must be registered LAST
         self.app.api_route("/{path:path}", methods=["GET", "POST", "PUT", "DELETE"])(self.proxy)
 
-    async def health_check(self, request: Request):
-        # TODO: do health check in background
-        pass
+    async def _start_background_health_check(self):
+        asyncio.create_task(self._health_check_loop())
+
+    async def _check_worker_health(self, url):
+        """Encapsulated health check logic for better maintainability."""
+        try:
+            response = await self.client.get(f"{url}/health", timeout=5.0)
+            if response.status_code == 200:
+                return url, True
+            logger.warning(f"[miles-router] Worker {url} is unhealthy (Status: {response.status_code})")
+        except Exception as e:
+            logger.warning(f"[miles-router] Worker {url} health check failed: {e}")
+        return url, False
+
+    async def _health_check_loop(self):
+        interval = self.args.rollout_health_check_interval
+
+        while True:
+            await asyncio.sleep(interval)
+
+            urls = list(self.worker_urls.keys())
+            if not urls:
+                continue
+
+            # Concurrent execution using a generator expression
+            results = await asyncio.gather(*(self._check_worker_health(url) for url in urls))
+
+            for url, is_healthy in results:
+                if not is_healthy:
+                    logger.warning(f"[miles-router] Removing unhealthy worker from pool: {url}")
+                    self.worker_urls.pop(url, None)
+
+            logger.info(f"[miles-router] Health check complete. {len(self.worker_urls)} workers active.")
 
     async def proxy(self, request: Request, path: str):
         """Proxy all other requests to the SGLang router"""