Sorry, brain fade. The question was somewhat rhetorical though. it was more 'how much current can the device draw before something on the board melts/shorts/otherwise fails?'.
We don't know what triggers the high draw - low current? current interruption? age? totally random component failure? none of the foregoing, it's something else?
We also don't know (but it may be found by experimentation) if the high draw is static, or if the draw increase is variable and increases with time/events (some of the above for example). We do know that my continuous draw was 100ma plus, yours 3.3ma (short period? but that number also outside specs), but we don't know why they differ. Given my unit was still operating, the board and its components can live at the 100-150ma range at least in the short term.