feat(watchdog): catch sluggish-device failures via soft-pattern counting

The container watchdog only restarted on three legacy "device clearly dead"
log lines, so today's failure mode (firmware briefly stalls and get_stats_*
/ get_battery commands time out with an empty error while passive RX
keeps working) never tripped it — leaving the user with 10-15 s freezes
several times a day and no automatic recovery.

DeviceManager now tracks two liveness signals:
- _last_rx_at, bumped on every RX_LOG_DATA event
- _consecutive_stats_failures, incremented on get_stats_* / get_bat
  exceptions and cleared on success

New /health/strict endpoint exposes these to the watchdog. It returns 503
when the device is connected but has 5+ consecutive stats failures, or
when no RX event has been seen for over 5 minutes on a serial transport.
The cheap /health endpoint keeps its lenient behavior so Docker's
healthcheck doesn't suddenly start tripping.

The watchdog's check_device_unresponsive() gains a "soft" pattern class
with a count threshold of 5 in the last 2 minutes — matching against
get_stats_core/radio/packets failed:, Failed to get battery:, and
Failed to get channel. Hard patterns still trigger on a single hit.

Deploy note: the watchdog runs as a host-level systemd service and is
NOT restarted by mcupdate, so after deploy run:
  sudo systemctl restart mc-webui-watchdog.service

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

scripts/watchdog/watchdog.py

@@ -469,23 +469,46 @@ def handle_unhealthy_container(container_name: str, status: dict):
 
 
 def check_device_unresponsive(container_name: str) -> bool:
-    """Check if the container logs indicate the USB device is unresponsive."""
+    """Check if the container logs indicate the USB device is unresponsive.
+
+    Two classes of patterns:
+      - HARD: any single occurrence triggers a restart. These are the
+        long-standing "device clearly dead" messages.
+      - SOFT: any of these failing >=5 times in the last 2 minutes triggers
+        a restart. Catches the "sluggish but not dead" mode (firmware
+        stalls on get_stats / get_bat commands while still answering
+        passive RX events).
+    """
     success, stdout, stderr = run_compose_command([
-        'logs', '--since', '1m', container_name
+        'logs', '--since', '2m', container_name
     ])
     if not success:
         return False
-        
-    error_patterns = [
+
+    hard_patterns = [
         "No response from meshcore node, disconnecting",
         "Device connected but self_info is empty",
-        "Failed to connect after 10 attempts"
+        "Failed to connect after 10 attempts",
     ]
-    
-    for pattern in error_patterns:
+
+    soft_patterns = [
+        "get_stats_core failed:",
+        "get_stats_radio failed:",
+        "get_stats_packets failed:",
+        "Failed to get battery:",
+        "Failed to get channel",
+    ]
+    SOFT_THRESHOLD = 5
+
+    for pattern in hard_patterns:
         if pattern in stdout:
             return True
-            
+
+    for pattern in soft_patterns:
+        if stdout.count(pattern) >= SOFT_THRESHOLD:
+            log(f"Soft-failure threshold tripped: '{pattern}' x{stdout.count(pattern)} in 2m", "WARN")
+            return True
+
     return False