Skip to content
TheThinkingHome edited this page Sep 2, 2026 · 5 revisions

Z-Wave

Device Sentinel recognizes Z-Wave, and every Z-Wave device is watched. The coordinator itself is not read yet.

What You Get

Each Z-Wave device has its own reporting rhythm learned, its freeze window set from that rhythm, and freeze, unavailable and unknown detection applied. Battery thresholds and falling-battery detection apply. Signal is recorded where a device exposes a signal entity. The problem list, the daily brief and every report cover Z-Wave devices.

What Is Not There Yet

The controller is not watched. If your Z-Wave controller stops, the devices behind it go quiet and each is judged on its own, so you get a row per device rather than one row naming the controller. On Zigbee2MQTT and ZHA that outage is recognized and the devices are cleared of blame.

No pairing-window recognition. When you include or replace a device by hand, turn on Maintenance Mode first. That tells Device Sentinel the recovery was yours, and the silence before it is set aside rather than learned as the device's normal rhythm. Without it, a hand-fix teaches a false rhythm.

No per-device second opinion. Freeze verdicts stand on Device Sentinel's own learning.

None of this changes whether a device fault is caught. It changes what happens when the controller itself is the problem.

Help Build Z-Wave Support

Z-Wave support needs someone who runs it. Reading a controller means knowing what it publishes when it starts, stops and returns, and what its devices' last-seen and signal entities look like, and none of that can be answered from the outside.

If you run Z-Wave and are willing to help:

  1. Download diagnostics: Settings, Devices and Services, Device Sentinel, the three-dot menu, Download diagnostics. The file carries device names and no passwords or tokens.
  2. Open an issue at Issues titled Z-Wave reference system and attach it.
  3. Say which controller you use, roughly how many devices you have, and your mains-versus-battery mix.

Expect a few small experiments after that, such as restarting the controller at an agreed moment and sending what the logs show.

See Also

Clone this wiki locally