ZHA

ZHA

Device Sentinel recognizes ZHA and reads its coordinator. Everything on this page is what you get on a ZHA system without configuring anything.

What You Get

Every device on ZHA is watched the same way as any other device in your house: its own reporting rhythm is learned, its freeze window is set from that rhythm, and freeze, unavailable and unknown detection all apply. 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 ZHA devices.

Your coordinator is watched too. A Bridge: ZHA sensor reads one of three states:

• running: the coordinator is up.
• down: the coordinator has stopped.
• unknown: nothing has been heard from it yet, which you may see briefly after a restart.

A coordinator outage does not become sixty device faults. When the ZHA coordinator goes down, the devices behind it fall silent through no fault of their own. You get one row naming the coordinator and how many devices sit behind it, one message, and everything clears when the coordinator returns. Silences that span the outage are set aside rather than learned, so an outage does not stretch every device's freeze window afterwards. Two devices keep their own rows: one already broken before the outage, and one that does not come back after it.

An outage is dated from when the radio stopped, not from when it was noticed, so the duration you read is the real one.

What ZHA Does Not Give You

No pairing-window recognition. Zigbee2MQTT publishes when its pairing window is open, so a device that returns during a re-pair is recognized as your hand rather than as the device's own rhythm. ZHA publishes nothing equivalent that can be relied on. When you re-pair or otherwise fix a ZHA 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.

No per-device second opinion. On Zigbee2MQTT the stack publishes its own opinion of each device's availability, shown beside Device Sentinel's freeze verdict as confirmation. ZHA has no equivalent reading, so the verdict stands on Device Sentinel's own learning.

Neither of these changes whether a fault is caught. They change how much corroboration you see beside it, and how a hand-fix is recognized.

If You Run a Large Mesh

One case is not yet covered from experience: a partial mesh failure, where one branch of the network dies while the coordinator stays up. On a small network every coordinator failure is total. If you run a large ZHA mesh and watch a branch die on its own, a diagnostics download taken while it is happening is worth more than any description of it. Open an issue titled "ZHA partial mesh" at Issues and attach it. The file carries device names and no passwords or tokens.

See Also

• The Bridge and Broker Sensors, for what each coordinator state means and where to see it.
• Maintenance Mode, for telling Device Sentinel that a recovery was your own hand.
• The Problem List, for what a coordinator outage looks like while it is running.