fix: power outage auto-recovery improvements

Mosquitto:
- Change restart: always (was unless-stopped) to survive daemon restarts
- Add healthcheck (pub/sub ping) for depends_on conditions
- All downstream services use condition: service_healthy

Party-lock:
- Add MQTT connection retry loop (was crashing on DNS race after restart)

Zigbee2mqtt:
- Add entrypoint wrapper that auto-applies BLZ bridge patch at startup
- Bridge patch no longer needs manual 'just z2m-patch-bridge' after image pulls

Mopidy:
- Add python3-gi + python3-gi-cairo via apt (fixes 'No module named gi')
- Set PYTHONPATH to expose system gi module to Python 3.12

infra/compose.yaml

@@ -2,7 +2,7 @@ services:
   mosquitto:
     image: eclipse-mosquitto:latest
     container_name: mosquitto
-    restart: unless-stopped
+    restart: always
     user: "1883:1883"
     ports:
       - "1883:1883"
@@ -10,6 +10,12 @@ services:
       - ./mosquitto/config:/mosquitto/config
       - ./mosquitto/data:/mosquitto/data
       - ./mosquitto/log:/mosquitto/log
+    healthcheck:
+      test: ["CMD", "sh", "-c", "timeout 3 mosquitto_sub -p 1883 -t healthcheck -C 1 --quiet & sleep 0.5 && mosquitto_pub -p 1883 -t healthcheck -m ok -r; wait"]
+      interval: 10s
+      timeout: 5s
+      retries: 3
+      start_period: 15s
 
   zigbee2mqtt:
     image: ghcr.io/pine64/zigbee2mqtt:latest-dev   # BLZ fork, NOT koenkk
@@ -23,6 +29,7 @@ services:
       - ./zigbee2mqtt:/app/data
     devices:
       - /dev/ttyUSB0:/dev/ttyUSB0
+    entrypoint: ["/app/data/scripts/entrypoint.sh"]
     environment:
       TZ: America/Los_Angeles
 
@@ -43,7 +50,8 @@ services:
     container_name: nodered
     restart: unless-stopped
     depends_on:
-      - mosquitto
+      mosquitto:
+        condition: service_healthy
     user: "1000:1000"
     ports:
       - "1880:1880"
@@ -65,7 +73,8 @@ services:
     container_name: doorbell-listener
     restart: unless-stopped
     depends_on:
-      - mosquitto
+      mosquitto:
+        condition: service_healthy
     volumes:
       - ./doorbell-listener/state:/data/doorbell-listener
     environment:
@@ -146,7 +155,8 @@ services:
     container_name: party-lock
     restart: unless-stopped
     depends_on:
-      - mosquitto
+      mosquitto:
+        condition: service_healthy
     volumes:
       - ./mosquitto/config:/mosquitto/config
       - /var/run/docker.sock:/var/run/docker.sock

infra/mopidy/Dockerfile

@@ -14,12 +14,20 @@ RUN apt-get update \
       libpango-1.0-0 \
       libpangoft2-1.0-0 \
       libgirepository-1.0-1 \
-      libgirepository1.0-dev \
       gir1.2-glib-2.0 \
       gir1.2-gstreamer-1.0 \
       gir1.2-gst-plugins-base-1.0 \
  && rm -rf /var/lib/apt/lists/*
 
+RUN apt-get update \
+ && apt-get install -y --no-install-recommends \
+      python3-gi \
+      python3-gi-cairo \
+ && rm -rf /var/lib/apt/lists/*
+
+# Expose the system Python's gi module to Python 3.12 (/usr/local).
+ENV PYTHONPATH=/usr/lib/python3/dist-packages:${PYTHONPATH}
+
 RUN pip install --no-cache-dir \
     "setuptools<80" \
     Mopidy==3.4.2 \

infra/party-lock/party-lock.py

@@ -51,7 +51,13 @@ def main():
     rebuild_acl()
     client = mqtt.Client(mqtt.CallbackAPIVersion.VERSION2)
     client.on_message = on_message
-    client.connect(MQTT_HOST, MQTT_PORT, 60)
+    while True:
+        try:
+            client.connect(MQTT_HOST, MQTT_PORT, 60)
+            break
+        except Exception as e:
+            print(f"mqtt connect failed: {e}; retrying in 5s", flush=True)
+            time.sleep(5)
     client.subscribe("party-lock/set", qos=1)
     client.loop_forever()
 

infra/zigbee2mqtt/scripts/entrypoint.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -e
+
+node /app/data/scripts/blz-bridge-patch.js
+
+exec docker-entrypoint.sh "$@"