Ultra-high-energy cosmic rays arrive carrying more energy in a single nucleus than any accelerator physics comfortably explains: events above 10^20 eV, tens of joules in one particle. Their existence imposes a brutal locality constraint, because protons and nuclei at these energies lose energy to collisions with CMB photons, the GZK effect, capping source distances at roughly 50 to 200 Mpc. Whatever makes them must be nearby, and nearby is mapped.
The map fails the requirement three ways. Energetics: accelerating particles to 10^20 eV demands source classes, AGN jets, gamma-ray bursts, starburst winds, operating at or beyond their theoretical Hilbert limits, and the local volume holds few plausible candidates. Anisotropy: nearby sources should imprint their sky positions on the arrival directions, yet the observed sky is broadly isotropic with only mild large-scale dipole structure, not the clustering a sparse local source population predicts (Aab et al. 2020). Composition: Auger data indicate increasingly heavy nuclei at the highest energies, which most candidate accelerators produce inefficiently and which propagation should fragment (Alves Batista et al. 2019). Sources energetic enough are too rare, sources common enough are too weak, and the particles themselves are the wrong species for the standard engines.
The standing is a mature mystery with mature instruments: the Auger and Telescope Array exposures keep growing, AugerPrime adds composition sensitivity event by event, and the source question remains open after six decades.