The observed scaling relations between the thermal Sunyaev-Zel'dovich (tSZ) signal and cluster mass deviate from the self-similar predictions of $\Lambda$CDM. Specifically, the $Y-M$ relation (integrated Compton parameter versus mass) shows a shallower slope and higher scatter than expected, and the normalization evolves with redshift in ways inconsistent with simple gravitational heating. These deviations suggest that non-gravitational processes (AGN feedback, cool-core disruption, or pre-heating) play a larger role than predicted, but the required feedback strengths are difficult to reconcile with observations of star formation rates and metal enrichment without fine-tuning (Planck Collaboration 2013; Battaglia et al. 2012).