Telemetry

Each registered device reports the following keys on every poll.

Key	Unit	Meaning
power	W	Instantaneous active power
voltage	V	RMS voltage
current	A	RMS current
powerFactor	-1..1	Power factor. Always 0 on the combined total, which has no single meaningful value
totalForwardEnergy	kWh	Cumulative energy consumed since the meter was last reset
totalReturnedEnergy	kWh	Cumulative energy exported to the grid
valid	boolean	false when the meter could not read the channel

Relay channels

A switch:N channel reports the same fields as a meter, plus its relay state:

Field	Unit	Notes
active	bool	Relay state. This is what the OpenBridge toggle controls.
power	W	Zero on non-metering relays.
voltage	V	Zero on non-metering relays.
current	A	Zero on non-metering relays.
totalForwardEnergy	kWh	Cumulative, counted by the device itself.

Non-metering relays report zeros rather than omitting the fields, so a channel's telemetry keys stay stable and history charts do not develop gaps if you swap hardware.

Units

Gen1 devices report cumulative energy in watt-hours. OpenBridge energy history is kept in kilowatt-hours, so the plugin divides by 1000 before reporting. A raw device value of 224585.1 becomes 224.5851 kWh.

Getting this wrong is silent and expensive: the history chart would read 1000x high and the error would only be obvious months later.

How the total is computed

The combined device is not simply a fourth channel:

• Power uses the device's own total_power when present, falling back to summing channels.
• Current and energy are summed across channels.
• Voltage is the mean of the valid channels. Summing three 240 V phases into 720 V would be meaningless.
• Power factor is reported as 0, since a single figure across phases with different loads would be misleading.

Energy history

OpenBridge samples totalForwardEnergy periodically into ~/.openbridge/energy-history/<deviceId>.json and serves day, month and year buckets from it. Because each phase is its own device, each gets its own history file and its own chart.