home-assistant-zigbee-esphome

Zigbee and ESPHome devices

Firmware updates

Use watch_firmware_update for device firmware — ESPHome, WLED, Zigbee, and anything else exposing an update.* entity. One call starts the update, monitors progress in real time, and reports the result:

watch_firmware_update(entity_id="update.device_firmware", start_update=true)

Do not poll get_states in a loop instead; this tool exists because that pattern burns context and misses the failure modes.

System updates (Core, OS, Supervisor, add-ons) are a different path:

1. get_available_updates()               -> what needs updating
2. update_component(component="core")    -> start, returns job_id
3. get_update_progress(job_id="...")     -> monitor

Ask before starting either. A firmware update can brick a device on a bad power supply, and a Core update takes the instance offline. Check get_backup_posture first when the update is a large one.

Inspecting a device

zigporter inspect "Device Name" --json (or zigporter inspect sensor.entity_id --json) is the one command that cross-references ZHA, Zigbee2MQTT and the Home Assistant registry at once — use it before concluding anything about a Zigbee device. zigporter list-devices --json inventories everything; zigporter list-z2m --json needs the Z2M URL configured in the add-on settings.

For the signal-quality picture, zigporter network-map --format table prints the mesh in the terminal, and zigporter network-map --output mesh.svg writes an SVG. Weak lqi, a device routing through a distant parent, or a router that has gone offline explain a large share of "the sensor keeps dropping out" reports.

zigporter check verifies connectivity before you trust any of the above.

Renaming — the thing zigporter is for

hab entity update and hab device update rename one thing and leave every reference to the old ID dangling. zigporter cascades: it patches automations, scripts, scenes and every Lovelace dashboard in one atomic pass.