Thanks for the info, Operahouse! The part I'm still not sure about is, how does it decide when the defrost cycle is complete? Does that depend on the timer (but not the compressor) running during the defrost? If that's the case, the behavior I saw (cycle never finished, frig getting way too warm meanwhile) could happen without any malfunction in the frig nor the inverter, simply because when the temp sensor cuts out the defrost heating element, the inverter goes to sleep.
And why do you say that "putting it on the same inverter as the freezer would be disasterous"? Or do you only mean that it would interfere with diagnosis? Other than diagnosis, it seem advantageous, since the freezer would turn on once in a while, causing the inverter to be on, which would at the same time advance the defrost cycle in the frig.