About
Connection Observer monitors smart home device connections by protocol and sends notifications when devices go offline. Whole integration families such as Zigbee, Z-Wave, Bluetooth, ESPHome and Shelly are selected for monitoring instead of individual entities, and labels can mark single entities to watch or ignore. Alerts can be sent immediately, as a scheduled summary on chosen days and times, or both, with an optional notice when a device reconnects. Alert delays can be set globally or per protocol, with recommended values available, and a cooldown, minimum offline duration and device or domain exclusions reduce noise. A sensor keeps recent disconnect events, notification text can be customized with templates, and an issue can appear in the Repairs panel after a set number of hours offline. It works locally, is set up in the UI, and provides binary sensors and sensors.
Automatic summary of the README and code of OleSint/ha-connection-observer. Something wrong, or are you the maintainer and want your own text here? Get in touch
Supports
Works locally Talks to your devices on your own network. No vendor cloud in between.
From the Marketplace Installed through the Marketplace, formerly HACS.
Actively maintained Its author released or pushed a change in the last six months.
Provides
- Sensors
Setting it up
Install it from the Marketplace and restart Home Assistant, then add it under Settings, Devices and services.
Connection Observer – Protocols
- Protocols to monitor
- Notification language
Connection Observer – Notifications
- Notification service(s)
- Notification target (optional)Only needed if your notify service sends to multiple recipients at once and you want to reach just one. If your service is notify.send_message, enter the specific recipient entity_id here (e.g. notify.smtp_jane_doe). For other services (e.g. telegram_bot.send_message), enter the raw target value that service expects instead (e.g. a chat ID).
- Immediate notification on disconnect
- Scheduled summary notification
- Summary time
- Summary days
- Notify on reconnect (opt-in)
Connection Observer – Test
- Send test notification now
Connection Observer – Advanced
- Global alert delay (minutes offline before an event is created)0 = instant. This is the default for all protocols unless overridden in the next step.
- Cooldown between immediate notifications per device (0 = no limit)
- Minimum offline duration to appear in summary (0 = all events)
- Startup re-check delay (minutes) before flagging an already-offline deviceAfter a restart, a device still unavailable is re-checked after this many minutes before being flagged offline. Increase this if you have several devices reconnecting simultaneously (e.g. many WiFi devices on one router).
- Include room / area in notifications
- Include manufacturer & model in notifications
- Excluded entity domains (e.g. sensor, button)
- Excluded devices (optional)
Connection Observer – Expert
- Apply recommended delays for all protocolsTick and click Submit to pre-fill all fields with the recommended values for each protocol. You can fine-tune the values afterwards.
- Zigbee (ZHA) – alert delayRecommended: 5 min | Zigbee mesh routing takes a moment
- Zigbee (deCONZ) – alert delayRecommended: 5 min | Zigbee mesh routing takes a moment
- Z-Wave (Z-Wave JS) – alert delayRecommended: 5 min | Z-Wave mesh routing takes a moment
- Matter – alert delayRecommended: 5 min | Mesh-like behaviour
- Thread (OTBR) – alert delayRecommended: 5 min | Thread mesh
- Bluetooth – alert delayRecommended: 10 min | BLE connection setup is slower
- BTHome – alert delayRecommended: 20 min | Passive BLE – rare advertisements
- RFXtrx (433 MHz) – alert delayRecommended: 10 min | One-way 433 MHz RF, no ACK
- MySensors – alert delayRecommended: 10 min | Slow polling
- Insteon – alert delayRecommended: 5 min | Proprietary bus, polling-based
- KNX – alert delayRecommended: 5 min | Wired bus, reliable but polled
- Velbus – alert delayRecommended: 5 min | Wired bus
- ESPHome – alert delayRecommended: 2 min | Persistent TCP, very fast detection
- Shelly – alert delayRecommended: 2 min | Persistent TCP, very fast detection
After setup
- Protocols to monitor
- Notification language
- Notification service(s)
- Notification target (optional)Only needed if your notify service sends to multiple recipients at once and you want to reach just one. If your service is notify.send_message, enter the specific recipient entity_id here (e.g. notify.smtp_jane_doe). For other services (e.g. telegram_bot.send_message), enter the raw target value that service expects instead (e.g. a chat ID).
- Immediate notification on disconnect
- Scheduled summary notification
- Summary time
- Summary days
- Notify on reconnect
- Global alert delay (minutes offline before an event is created)0 = instant. This is the default for all protocols unless overridden in the Expert step.
- Cooldown between immediate notifications per device (0 = no limit)
- Minimum offline duration for summary (minutes, 0 = all events)
- Startup re-check delay (minutes) before flagging an already-offline deviceAfter a restart, a device still unavailable is re-checked after this many minutes before being flagged offline. Increase this if you have several devices reconnecting simultaneously (e.g. many WiFi devices on one router).
- Include room / area in notifications
- Include manufacturer & model in notifications
- Excluded entity domains (e.g. sensor, button)
- Excluded devices
- HA Repairs issue after N hours offline (0 = disabled)
- Apply recommended delays for all protocolsTick and click Submit to pre-fill all fields with the recommended values for each protocol. You can fine-tune the values afterwards.
- Zigbee (ZHA) – alert delayRecommended: 5 min | Zigbee mesh routing takes a moment
- Zigbee (deCONZ) – alert delayRecommended: 5 min | Zigbee mesh routing takes a moment
- Z-Wave (Z-Wave JS) – alert delayRecommended: 5 min | Z-Wave mesh routing takes a moment
- Matter – alert delayRecommended: 5 min | Mesh-like behaviour
- Thread (OTBR) – alert delayRecommended: 5 min | Thread mesh
- Bluetooth – alert delayRecommended: 10 min | BLE connection setup is slower
- BTHome – alert delayRecommended: 20 min | Passive BLE – rare advertisements
- RFXtrx (433 MHz) – alert delayRecommended: 10 min | One-way 433 MHz RF, no ACK
- MySensors – alert delayRecommended: 10 min | Slow polling
- Insteon – alert delayRecommended: 5 min | Proprietary bus, polling-based
- KNX – alert delayRecommended: 5 min | Wired bus, reliable but polled
- Velbus – alert delayRecommended: 5 min | Wired bus
- ESPHome – alert delayRecommended: 2 min | Persistent TCP, very fast detection
- Shelly – alert delayRecommended: 2 min | Persistent TCP, very fast detection
- Immediate notification – title (empty = default)Variables: device_name, protocol, time, area, model
- Immediate notification – message (empty = default)Variables: device_name, protocol, time, area, model
- Reconnect notification – title (empty = default)Variables: device_name
- Reconnect notification – message (empty = default)Variables: device_name
- Summary – title (empty = default)Variables: count
- Summary line – resolved event (empty = default)Variables: device_name, area, protocol, time_offline, time_online
- Summary line – ongoing offline (empty = default)Variables: device_name, area, protocol, time_offline
- Send test notification now
Texts from the code of OleSint/ha-connection-observer.
From the README
- Protocol-based monitoring – select entire integration families (Zigbee, Z-Wave, Hue, ESPHome, Shelly, Sonos, …) instead of configuring individual entities. 100+ integrations supported out of the box.
- Immediate notifications – get alerted the moment a device goes offline
- Scheduled summaries – receive a digest on configurable days and times listing everything that disconnected since the last summary, including whether the device came back online and when
- Both modes simultaneously – immediate + summary can both be active at once; summary is pre-selected by default
- Reconnect notifications – opt-in alert when a device comes back online
- Alert delay – opt-in: only create an event after a device has been offline for N minutes, filtering out brief dropouts entirely
- Cooldown – opt-in: limit immediate notifications per device to once every N minutes
- Minimum offline duration – opt-in: suppress short blips from the summary without losing them
- Room / area – opt-in: include the HA area name in notifications
- Manufacturer & model – opt-in: include device info in immediate notifications
- Event history sensor (v1.3.0) – sensor.event_history exposes the last 100 disconnect events as a structured events attribute with timestamps, area, protocol, and reconnect info — ready for…
- Label-based monitoring (v1.2.0) – assign observer_critical, observer_watch, or observer_ignore labels to any HA entity; works independently of protocols, no config-flow changes needed after setup
Shows in Vulpo
Chart card, Apple Watch.Information
- Installs
- 96 sharing analytics
- GitHub stars
- 15
- Connection
- Works locally
- Upkeep
- Updated in the last 6 months
- Version
- v1.3.13.1,
- Category
- Messaging and notifications, Push notifications
- Kind
- Marketplace integration
- Needs
- Needs the Marketplace (HACS on 2026.10 and older)
- Domain
connection_observer- Issues
- Report an issue









