My belief, based on your sensor testing, is that the gauge may be wonky. Trying the digital gauge might prove that, although the temperature signal appears that it may be retransmitted by the cluster, so that would also mean the cluster could be at fault if that proves to be true, and both gauges would read inaccurately. A dirty, loose, or bad connection which changed the resistance anywhere in the circuit could also be at fault, as that could change both the gauge readings and the fan response and warnings.
I still feel that your fan is not operating properly, although if it came on and stayed on during your testing it is probably OK and your Spyder temperature and fan are cycling as they normally do. If the fan comes on then cycles off, but the temp continues to rise, there is a problem. If the fan comes on, the temp gauge drops slightly, and it cycles back off, that would be normal. A fan that runs continuously indicates running hotter than normal, but not necessarily overheating. The fan should come on around the sixth bar or mark, and cycle off around 5 marks or bars. Be aware that the viewing angle may change the perception of where it comes on with the analog gauge. Overheating warnings should not wait until the top of the gauge, but come on at the bottom of the red or even nearer seven bars. Your testing method may have had something to do with that. A direct heat source is pretty uncontrollable, reacts fast, and could heat the sensor metal so that the temperature continued to rise after the heat was removed.
Boiling water in a pot should not be enough to trigger a warning...water under pressure and with coolant added boils at a higher temperature than plain water under atmospheric pressure. Your test went as I would expect it to go, in that regard.