Analogue components tend to fail consistently once they've gone they've gone. (To paraphrase an advertisement).
Their degradation is usually permanent leading to consistent rather than intermittent failure, although initial signs may be intermittent for a time the failure usually hardens up fairly quickly.
Failure also tends to be at the very beginning of the components life or after a considerable time, the so called bathtub graph of failure.
Although as you say the components may be of poor quality hastening the overall speed of onset of failure.
The DSM2 system can be prone to interference if the two channels chosen by the software at start-up are very close (re: Bruce's video about DSM2).
But it's really difficult to detect the presence of transitory interference on 2.4 GHz without expensive exotic equipment. So interference can not be eliminated as a suspect!
My own crash with DSM2 was due to an ESC failure brownout resulting in a landing in a tree top as I regained partial control.
Running up the motor on the ground after recovery resulted in an impressive ESC explosion!
I always use an independent switching BEC now and advocate them to everyone.
I have experienced some weird glitches and lockouts on DSM2 with the standard RX's but I think they were probably brownouts due to using the ESC's regulators.
Interesting that the 35 Meg gear worked so faultlessly, probably because it's trendy to go 2.4 and the band is relatively clear now, we should probably still use it for internal combustion engines.
Compromising with out of control government is like living with a lion, sooner or later the bloody thing will eat you .