feat: monitor heartbeat + background poller skips active-monitored units

- Heartbeat: if nothing has been broadcast in MONITOR_HEARTBEAT_S (default
  25s) — e.g. device offline and silent — send a non-cached keepalive frame
  so a reverse proxy (NPM) doesn't drop the idle WS. New subscribers still
  get the last real frame, not a heartbeat.
- Poller-skip: the 60s background poller now skips any unit with a running
  monitor (MonitorManager.is_active). The monitor already polls it ~1Hz and
  keeps the status cache fresh, so the background poll was redundant and just
  added load/lock-contention on the device's single connection (and churn,
  which matters for the cellular wedge). Trade-off: the FTP start-time sync
  (only in the poller) doesn't run while a unit is actively monitored — fine,
  since reports take the authoritative start time from the FTP .rnd data.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

app/monitor.py

@@ -31,6 +31,10 @@ logger = logging.getLogger(__name__)
 # poll) already paces the effective rate to a few seconds; this is the extra idle.
 MONITOR_POLL_INTERVAL = float(os.getenv("MONITOR_POLL_INTERVAL", "1.0"))
 
+# If nothing has been broadcast in this many seconds (e.g. device offline and
+# silent), send a keepalive frame so reverse proxies don't drop the idle WS.
+MONITOR_HEARTBEAT_S = float(os.getenv("MONITOR_HEARTBEAT_S", "25"))
+
 
 def _snapshot_payload(snap, unit_id: str, measurement_start_time) -> dict:
     """Build the broadcast payload — same shape as the DRD stream, but DOD-sourced
@@ -107,6 +111,8 @@ class DeviceMonitor:
 
     async def _run(self) -> None:
         logger.info(f"[MONITOR] {self.unit_id}: feed started")
+        loop = asyncio.get_running_loop()
+        last_send = loop.time()
         try:
             while self._has_demand():
                 snap, mst = await self._poll_once()
@@ -116,6 +122,7 @@ class DeviceMonitor:
                     payload = _snapshot_payload(snap, self.unit_id, mst)
                     payload["feed_status"] = "ok"
                     self._broadcast(payload)
+                    last_send = loop.time()
                     try:
                         await alert_evaluator.evaluate(self.unit_id, snap)
                     except Exception as e:
@@ -131,6 +138,20 @@ class DeviceMonitor:
                             "timestamp": datetime.utcnow().isoformat(),
                             "feed_status": "unreachable",
                         })
+                        last_send = loop.time()
+
+                # Heartbeat: during quiet/offline stretches, send a keepalive so an
+                # idle WS isn't dropped by a reverse proxy. Not cached (new subscribers
+                # should still get the last real frame, not a heartbeat).
+                if loop.time() - last_send >= MONITOR_HEARTBEAT_S:
+                    self._broadcast({
+                        "unit_id": self.unit_id,
+                        "timestamp": datetime.utcnow().isoformat(),
+                        "feed_status": "ok" if self._reachable else "unreachable",
+                        "heartbeat": True,
+                    }, cache=False)
+                    last_send = loop.time()
+
                 await asyncio.sleep(MONITOR_POLL_INTERVAL)
         finally:
             logger.info(f"[MONITOR] {self.unit_id}: feed stopped")
@@ -161,8 +182,9 @@ class DeviceMonitor:
         finally:
             db.close()
 
-    def _broadcast(self, payload: dict) -> None:
-        self._last_payload = payload  # cache for replay to new subscribers
+    def _broadcast(self, payload: dict, cache: bool = True) -> None:
+        if cache:
+            self._last_payload = payload  # replayed to new subscribers
         for q in list(self._subscribers):
             try:
                 q.put_nowait(payload)
@@ -186,6 +208,12 @@ class MonitorManager:
                 self._monitors[unit_id] = m
             return m
 
+    def is_active(self, unit_id: str) -> bool:
+        """True if this unit has a running monitor feed (so the background poller
+        can skip it — the monitor already polls it more often)."""
+        m = self._monitors.get(unit_id)
+        return m is not None and m.running
+
     def status(self) -> dict:
         return {
             uid: {