Sensor nodes equipped with renewable energy sources are capable of recharging their batteries and supporting data collection and transmission indefinitely. Energy and data management of these types of systems is challenging primarily due to the variability of renewable energy sources and transmission channels. This paper explores a methodology for designing the control of such systems. The goal is to jointly control the energy usage and data sampling rate to maximize the long-term performance of the system subject to the constraints imposed by the available energy and data. The design of this control is based on estimates of the large deviations of the energy stored in the battery and of the queued data. A low-complexity control policy is proposed that does not depend on the instantaneous charge of the battery and data backlog and almost maximizes the long-term data transmission rate. Moreover, the results show that one can decouple the analysis of the energy and of the data queue without much loss in performance.