The availability of powerful smart devices capable of supporting ubiquitous Internet access, and the roll-out of advanced networks such as 4G and 5G capable of guaranteeing Quality of Service (QoS), has led to an exponential increase in demand for scarce network resources. The increase in number of users can result in an increased revenue for the mobile network operators. However, unlike wired networks for which capacity can be increased easily, the capacity for cellular networks is limited. This kind of paradox, i.e., rapid growth in customer demand for mobile data services on one hand, and the significant challenges faced by operators for pricing and allocation of the limited resources on the other, is a challenging issue. As a result, cellular network operators rely on techniques such as traffic shaping and zero-rating to manage user demands. In the case of zero-rating, network operator generally limits the throughput provided to certain content or services, and the user generated traffic for those services is not counted in their monthly bill. This not only helps to alleviate congestion but also to expand internet access to those who cannot afford data plans. We propose a framework whereby zero-rating services can be limited by overall QoS rather than only by the data rates. The proposed approach is flexible in that it applies to services with constraints on throughput, delay, error-rate and jitter as well as any combination of these constraints. The approach is also simple since it uses information that is already available to the base-station scheduler and so can be easily deployed.