fix(health): move Docker client+ping off the event loop

Under CPU saturation the sync docker.from_env()/ping() calls could miss
their socket timeout, cache _docker_healthy=False, and return 503 for
the full 5s TTL window. Both calls now run on a thread so the event
loop keeps serving other requests while Docker is being probed.

decnet/web/router/health/api_get_health.py

@@ -102,8 +102,8 @@ async def get_health(user: dict = Depends(require_viewer)) -> Any:
             import docker
 
             if _docker_client is None:
-                _docker_client = docker.from_env()
+                _docker_client = await asyncio.to_thread(docker.from_env)
-            _docker_client.ping()
+            await asyncio.to_thread(_docker_client.ping)
             _docker_healthy = True
             _docker_detail = ""
         except Exception as exc: