feat: Gateway health monitoring (Hermes CLI, API, frontend)

- Hermes CLI wrapper using subprocess.run for gateway status
- Gateway API: status list (with filter), events (placeholder), start/stop/restart with audit logging
- Frontend GatewayPage: timeline, filter bar, status table with 10s polling, operation modals
- usePolling hook for periodic data refresh

Co-Authored-By: Claude <noreply@anthropic.com>

backend/app/api/gateway.py

@@ -0,0 +1,120 @@
+"""Gateway API — 状态、事件、启停操作。"""
+import logging
+import subprocess
+
+from fastapi import APIRouter, Depends, HTTPException, Query
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.config import settings
+from app.database import get_db
+from app.middleware.auth import require_auth, require_role
+from app.models.audit import AuditLog
+from app.models.user import User
+from app.services import hermes_cli
+
+logger = logging.getLogger(__name__)
+
+router = APIRouter(prefix="/gateway", tags=["gateway"])
+
+# ── 占位事件数据（TODO: 从 hermes 日志或系统日志解析） ──
+
+PLACEHOLDER_EVENTS = [
+    {"time": "2026-07-22T18:45", "profile": "monitor", "type": "heartbeat_timeout",
+     "detail": "连续 3 次心跳无响应"},
+    {"time": "2026-07-22T14:20", "profile": "feishu-gw", "type": "feishu_disconnect",
+     "detail": "WebSocket 断开 8 分钟"},
+    {"time": "2026-07-22T10:30", "profile": "gateway-monitor", "type": "restart",
+     "detail": "Gateway 重启"},
+]
+
+
+@router.get("/status")
+async def get_gateway_status(
+    status: str | None = Query(None, pattern="^(online|offline|warn)$"),
+    _user: User = Depends(require_auth),
+):
+    """返回全部 Profile 的 Gateway 状态列表，支持 ?status= 筛选。"""
+    gateways = hermes_cli.gateway_status_all()
+    for gw in gateways:
+        gw.setdefault("downtime_minutes", 0)
+        gw.setdefault("restarts_today", 0)
+        gw.setdefault("last_event", "")
+        gw.setdefault("profile_type", "worker")
+    if status:
+        gateways = [g for g in gateways if g.get("status") == status]
+    return gateways
+
+
+@router.get("/events")
+async def get_gateway_events(
+    _user: User = Depends(require_auth),
+):
+    """返回最近 24h 异常事件。"""
+    # TODO: 从 hermes CLI 日志或系统日志解析真实事件
+    return PLACEHOLDER_EVENTS
+
+
+async def _operate_gateway(
+    profile: str,
+    action: str,
+    user: User,
+    db: AsyncSession,
+) -> dict:
+    """执行 hermes gateway {action} --profile {profile} 并写入审计日志。"""
+    try:
+        proc = subprocess.run(
+            [settings.hermes_cli_path, "gateway", action, "--profile", profile],
+            capture_output=True,
+            text=True,
+            timeout=30,
+        )
+        success = proc.returncode == 0
+        detail = (proc.stdout or proc.stderr).strip()
+    except Exception as e:
+        success = False
+        detail = str(e)
+
+    audit = AuditLog(
+        user_id=user.id,
+        username=user.username,
+        action=f"gateway:{action}",
+        target=f"profile:{profile}",
+        detail=detail,
+    )
+    db.add(audit)
+    await db.commit()
+
+    if not success:
+        raise HTTPException(status_code=500, detail=f"Gateway {action} 失败: {detail}")
+
+    return {"success": True, "action": action, "profile": profile, "detail": detail}
+
+
+@router.post("/{profile}/start")
+async def start_gateway(
+    profile: str,
+    user: User = Depends(require_role("owner", "operator")),
+    db: AsyncSession = Depends(get_db),
+):
+    """启动 Gateway。"""
+    return await _operate_gateway(profile, "start", user, db)
+
+
+@router.post("/{profile}/stop")
+async def stop_gateway(
+    profile: str,
+    user: User = Depends(require_role("owner", "operator")),
+    db: AsyncSession = Depends(get_db),
+):
+    """停止 Gateway。"""
+    return await _operate_gateway(profile, "stop", user, db)
+
+
+@router.post("/{profile}/restart")
+async def restart_gateway(
+    profile: str,
+    user: User = Depends(require_role("owner", "operator")),
+    db: AsyncSession = Depends(get_db),
+):
+    """重启 Gateway。"""
+    return await _operate_gateway(profile, "restart", user, db)

backend/app/main.py

@@ -4,7 +4,7 @@ from fastapi import FastAPI
 from fastapi.middleware.cors import CORSMiddleware
 
 from app.database import engine, Base
-from app.api import auth, users, audit, dashboard
+from app.api import auth, users, audit, dashboard, gateway
 from app.middleware.audit import AuditMiddleware
 
 
@@ -35,6 +35,7 @@ app.include_router(auth.router)
 app.include_router(users.router)
 app.include_router(audit.router)
 app.include_router(dashboard.router, prefix="/api/v1")
+app.include_router(gateway.router, prefix="/api/v1")
 
 
 @app.get("/api/v1/health")

backend/app/services/hermes_cli.py

@@ -0,0 +1,43 @@
+"""Hermes CLI 封装 — 通过 subprocess 调用 hermes 命令。"""
+import json
+import logging
+import subprocess
+
+from app.config import settings
+
+logger = logging.getLogger(__name__)
+
+
+def _run_hermes(*args: str, timeout: int = 15) -> subprocess.CompletedProcess:
+    """执行 hermes CLI 命令，返回 CompletedProcess。"""
+    return subprocess.run(
+        [settings.hermes_cli_path, *args],
+        capture_output=True,
+        text=True,
+        timeout=timeout,
+    )
+
+
+def gateway_status_all() -> list[dict]:
+    """执行 `hermes gateway status`，返回解析后的 Gateway 状态列表。
+
+    出错时返回空列表并记录告警日志。
+    """
+    try:
+        proc = _run_hermes("gateway", "status")
+        if proc.returncode != 0:
+            logger.warning("hermes gateway status 返回非零: %s", proc.stderr.strip())
+            return []
+        data = json.loads(proc.stdout)
+        if isinstance(data, list):
+            return data
+        return data.get("gateways", [])
+    except FileNotFoundError:
+        logger.warning("Hermes CLI 未找到: %s", settings.hermes_cli_path)
+    except json.JSONDecodeError as e:
+        logger.warning("解析 hermes gateway status 输出失败: %s", e)
+    except subprocess.TimeoutExpired:
+        logger.warning("hermes gateway status 执行超时")
+    except Exception as e:
+        logger.warning("调用 hermes gateway status 异常: %s", e)
+    return []