feat: per-service status checks with offline colouring

Adds optional live status checking per service (not just per node),
requested as a follow-up to issue #196.

Backend:
- New check_service / check_services: HTTP(S) GET for web services, TCP
  connect otherwise; UDP and port-less non-web services stay 'unknown'.
- New scheduler job 'service_checks', independent interval (default 300s),
  added/removed live via set_service_checks_enabled.
- Settings gain service_check_enabled + service_check_interval (>=30s),
  persisted to scan_config.json. New WS message type 'service_status'.

Frontend:
- Live per-service status overlay in canvasStore (not persisted, so it
  never round-trips through canvas save), fed by the WS message.
- DetailPanel + canvas node service rows: offline service turns red
  (#f85149), otherwise keeps its category colour.
- SettingsModal: toggle + interval input (default 300s / 5 min).

Off by default — no behaviour change until enabled.

ha-relevant: yes

backend/app/api/routes/settings.py

@@ -1,20 +1,27 @@
 """App-level settings (status checker interval, etc.)."""
 from fastapi import APIRouter, Depends, HTTPException
-from pydantic import BaseModel
+from pydantic import BaseModel, Field
 
 from app.api.deps import get_current_user
 from app.core.config import settings
+from app.core.scheduler import reschedule_service_checks, set_service_checks_enabled
 
 router = APIRouter()
 
 
 class AppSettings(BaseModel):
     interval_seconds: int
+    service_check_enabled: bool = False
+    service_check_interval: int = Field(default=300, ge=30)
 
 
 @router.get("", response_model=AppSettings)
 async def get_settings(_: str = Depends(get_current_user)) -> AppSettings:
-    return AppSettings(interval_seconds=settings.status_checker_interval)
+    return AppSettings(
+        interval_seconds=settings.status_checker_interval,
+        service_check_enabled=settings.service_check_enabled,
+        service_check_interval=settings.service_check_interval,
+    )
 
 
 @router.post("", response_model=AppSettings)
@@ -23,7 +30,13 @@ async def update_settings(
 ) -> AppSettings:
     try:
         settings.status_checker_interval = payload.interval_seconds
+        settings.service_check_enabled = payload.service_check_enabled
+        settings.service_check_interval = payload.service_check_interval
         settings.save_overrides()
+        # Apply the service-check schedule live.
+        set_service_checks_enabled(payload.service_check_enabled)
+        if payload.service_check_enabled:
+            reschedule_service_checks(payload.service_check_interval)
         return payload
     except Exception as exc:
         raise HTTPException(status_code=500, detail=str(exc)) from exc

backend/app/api/routes/status.py

@@ -54,6 +54,15 @@ async def broadcast_status(node_id: str, status: str, checked_at: str, response_
     }))
 
 
+async def broadcast_service_status(node_id: str, services: list[dict], checked_at: str) -> None:
+    await _broadcast(json.dumps({
+        "type": "service_status",
+        "node_id": node_id,
+        "services": services,
+        "checked_at": checked_at,
+    }))
+
+
 async def broadcast_scan_update(run_id: str, devices_found: int) -> None:
     await _broadcast(json.dumps({
         "type": "scan_device_found",

backend/app/core/config.py

@@ -51,6 +51,10 @@ class Settings(BaseSettings):
     # Status checker
     status_checker_interval: int = 60
 
+    # Per-service status checker (independent of node checks). Off by default.
+    service_check_enabled: bool = False
+    service_check_interval: int = 300
+
     # MCP service key — set MCP_SERVICE_KEY in .env
     # Used by the MCP server to authenticate against the backend without a user password.
     # Leave empty to disable MCP service key auth.
@@ -77,6 +81,10 @@ class Settings(BaseSettings):
                 self.scanner_ranges = data["scanner_ranges"]
             if "status_checker_interval" in data:
                 self.status_checker_interval = int(data["status_checker_interval"])
+            if "service_check_enabled" in data:
+                self.service_check_enabled = bool(data["service_check_enabled"])
+            if "service_check_interval" in data:
+                self.service_check_interval = int(data["service_check_interval"])
         except Exception:
             pass
 
@@ -86,6 +94,8 @@ class Settings(BaseSettings):
         self._override_path().write_text(json.dumps({
             "scanner_ranges": self.scanner_ranges,
             "status_checker_interval": self.status_checker_interval,
+            "service_check_enabled": self.service_check_enabled,
+            "service_check_interval": self.service_check_interval,
         }))
 
 

backend/app/core/scheduler.py

@@ -9,7 +9,7 @@ from sqlalchemy import select
 from app.core.config import settings
 from app.db.database import AsyncSessionLocal
 from app.db.models import Node
-from app.services.status_checker import check_node
+from app.services.status_checker import check_node, check_services
 
 logger = logging.getLogger(__name__)
 
@@ -73,6 +73,50 @@ async def _run_status_checks() -> None:
     ])
 
 
+def _node_host(ip: str | None, hostname: str | None) -> str | None:
+    """Pick the address to probe services on: first IP, else hostname."""
+    if ip:
+        first = ip.split(",")[0].strip()
+        if first:
+            return first
+    return hostname or None
+
+
+async def _run_service_checks() -> None:
+    """Check every service of every node and broadcast per-service results."""
+    if not settings.service_check_enabled:
+        return
+    from app.api.routes.status import broadcast_service_status  # avoid circular import
+
+    async with AsyncSessionLocal() as db:
+        result = await db.execute(select(Node))
+        nodes = result.scalars().all()
+        checkable = [
+            (n.id, _node_host(n.ip, n.hostname), list(n.services or []))
+            for n in nodes
+            if n.services
+        ]
+
+    now = datetime.now(timezone.utc).isoformat()
+    for node_id, host, services in checkable:
+        try:
+            statuses = await check_services(host, services)
+            await broadcast_service_status(node_id=node_id, services=statuses, checked_at=now)
+        except Exception as exc:
+            logger.error("Service checks failed for node %s: %s", node_id, exc)
+
+
+def _add_service_check_job() -> None:
+    scheduler.add_job(
+        _run_service_checks,
+        "interval",
+        seconds=settings.service_check_interval,
+        id="service_checks",
+        max_instances=1,
+        coalesce=True,
+    )
+
+
 def start_scheduler() -> None:
     global scheduler
     if scheduler.running:
@@ -89,6 +133,8 @@ def start_scheduler() -> None:
         max_instances=1,
         coalesce=True,
     )
+    if settings.service_check_enabled:
+        _add_service_check_job()
     scheduler.start()
     logger.info("Scheduler started — status checks every %ds", settings.status_checker_interval)
 
@@ -104,6 +150,31 @@ def reschedule_status_checks(interval_seconds: int) -> None:
     logger.info("Status checks rescheduled to every %ds", interval_seconds)
 
 
+def reschedule_service_checks(interval_seconds: int) -> None:
+    """Update the service-check interval on the running scheduler (if enabled)."""
+    if interval_seconds < 30:
+        raise ValueError(f"interval_seconds must be >= 30, got {interval_seconds}")
+    if not scheduler.running:
+        logger.warning("Scheduler not running, skipping reschedule")
+        return
+    if scheduler.get_job("service_checks"):
+        scheduler.reschedule_job("service_checks", trigger="interval", seconds=interval_seconds)
+        logger.info("Service checks rescheduled to every %ds", interval_seconds)
+
+
+def set_service_checks_enabled(enabled: bool) -> None:
+    """Add or remove the service-check job on the running scheduler."""
+    if not scheduler.running:
+        return
+    job = scheduler.get_job("service_checks")
+    if enabled and not job:
+        _add_service_check_job()
+        logger.info("Service checks enabled — every %ds", settings.service_check_interval)
+    elif not enabled and job:
+        scheduler.remove_job("service_checks")
+        logger.info("Service checks disabled")
+
+
 def stop_scheduler() -> None:
     if scheduler.running:
         scheduler.shutdown(wait=False)

backend/app/services/status_checker.py

@@ -118,3 +118,74 @@ async def _tcp_connect(host: str, port: int) -> bool:
         return True
     except (TimeoutError, OSError, socket.gaierror):
         return False
+
+
+# --- Per-service status checks ---
+
+# Ports that are definitely not HTTP/web — mirror of frontend serviceUrl.ts.
+# SSH (22) is handled as a plain TCP check, so it is intentionally absent here.
+_NON_HTTP_PORTS = frozenset({
+    21, 23, 25, 465, 587, 53, 110, 143, 993, 995, 389, 636, 445, 514,
+    1433, 3306, 5432, 5672, 6379, 9092, 11211, 27017, 27018,
+})
+_HTTPS_PORTS = frozenset({443, 8443})
+
+
+def _service_host(svc: dict[str, Any], host: str) -> str:
+    """Bracket bare IPv6 literals for use in a URL."""
+    return f"[{host}]" if _is_ipv6(host) else host
+
+
+async def check_service(svc: dict[str, Any], host: str | None) -> str:
+    """Check a single service. Returns 'online' | 'offline' | 'unknown'.
+
+    Web services get an HTTP(S) GET; everything else with a port gets a TCP
+    connect. UDP services and port-less non-web services are 'unknown' so they
+    keep their category colour rather than flashing red.
+    """
+    if not host or host.startswith("-"):
+        return "unknown"
+    if str(svc.get("protocol", "")).lower() == "udp":
+        return "unknown"
+
+    port = svc.get("port")
+    port = int(port) if isinstance(port, int) or (isinstance(port, str) and port.isdigit()) else None
+
+    try:
+        if port is not None and port in _NON_HTTP_PORTS:
+            return "online" if await _tcp_connect(host, port) else "offline"
+
+        name = str(svc.get("service_name", "")).lower()
+        is_web = port is None or port not in _NON_HTTP_PORTS
+        if is_web and (port is not None or "http" in name):
+            scheme = "https" if (
+                port in _HTTPS_PORTS or "https" in name or "ssl" in name or "tls" in name
+            ) else "http"
+            url_host = _service_host(svc, host)
+            url = f"{scheme}://{url_host}" + (f":{port}" if port is not None else "")
+            return "online" if await _http_get(url, verify=False) else "offline"
+
+        if port is not None:
+            return "online" if await _tcp_connect(host, port) else "offline"
+
+        return "unknown"
+    except Exception as exc:
+        logger.debug("Service check failed for %s:%s (%s)", host, port, exc)
+        return "offline"
+
+
+async def check_services(
+    host: str | None, services: list[dict[str, Any]], concurrency: int = 10
+) -> list[dict[str, Any]]:
+    """Check every service against host concurrently (bounded).
+
+    Returns a list of {port, protocol, status} dicts, one per input service.
+    """
+    sem = asyncio.Semaphore(concurrency)
+
+    async def _one(svc: dict[str, Any]) -> dict[str, Any]:
+        async with sem:
+            status = await check_service(svc, host)
+        return {"port": svc.get("port"), "protocol": svc.get("protocol"), "status": status}
+
+    return await asyncio.gather(*[_one(s) for s in services]) if services else []