Millimeter‐scale rock particles called chondrules are the principal components of the most common meteorites, chondrites. Hence, chondrules were arguably the most abundant components of the early solar system at the time of planetesimal accretion. Despite their fundamental importance, the existence of chondrules would not be predicted from current observations and models of young planetary systems. There are many different models for chondrule formation, but no single model satisfies the many constraints determined from their mineralogical and chemical properties and from chondrule analog experiments. Significant recent progress has shown that several models can satisfy first‐order constraints and successfully reproduce chondrule thermal histories. However, second‐ and third‐order constraints such as chondrule size ranges, open system behavior, oxidation states, reheating, and chemical diversity have not generally been addressed. Chondrule formation models include those based on processes that are known to occur in protoplanetary disk environments, including interactions with the early active Sun, impacts and collisions between planetary bodies, and radiative heating. Other models for chondrule heating mechanisms are based on hypothetical processes that are possible but have not been observed, like shock waves, planetesimal bow shocks, and lightning. We examine the evidence for the canonical view of chondrule formation, in which chondrules were free‐floating particles in the protoplanetary disk, and the noncanonical view, in which chondrules were the by‐products of planetesimal formation. The fundamental difference between these approaches has a bearing on the importance of chondrules during planet formation and the relevance of chondrules to interpreting the evolution of protoplanetary disks and planetary systems.