1. An examination of DNA might reveal that gibbons share more in common with fir trees than fruit bats. That would be quite a problem... OR:
2. Fossils of dinosaurs in Quaternary sediments, with no other explanation for their arrival.
3. Perfectly preserved human remains dated to 1.2 billion years ago.
Any of those, or a hundred million variations on them, would pose a serious hurdle for evolution.
Yes. Still rather extremist.
Anything descrepancies that might ralistically be observed would cause a modification of theory rather than dumping it.