This article does a good job of explaining how failure modes/alternative usages can make that definition fuzzy: https://www.csemag.com/articles/understanding-overcurrent-pr...

Practically speaking, any intermittent device that _can_ run continuously or can fail to a continuous needs to be considered as such for safety purposes.

For example, a fridge should only run the compressor intermittently, but it has two obvious cases where the compressor could run indefinitely:

* An influx of heat, like filling an empty fridge with room temperature cans.

* A door being left open.

In the case of the kettle, it will likely be evaluated against it's "nominal" draw after the initial startup. If the auto-off sensor were to fail, it could run continuously at the tempurature.

I probably can’t speak to your jurisdiction, but no electrical inspector I’ve ever met would consider those loads continuous - in its precise, non-colloquial meaning - for the purpose of rating the circuit.

An appliance manufacturer may be under different UL regulations for failure modes of a device on a specific rated circuit. I don’t know anything about that.

[delayed]

If this is the standard then everything is a continuous load. A soldering iron or a hair dryer or a blender might get left on forever, who knows. Maybe I want the microwave to run 24 hours a day. If everything is a continuous load, why would the NEC make a difference at all?

Even in the context of your refrigerator example (which I don't necessarily agree with), a kettle should have two thermostats - a primary one that keeps the temperature setpoint, and a safety cutoff in case the primary stops working. In both fail and the kettle is somehow drawing its nominal power for hours at a time, the main safety hazard is going to be the kettle itself trying to dissipate >1Kw.