Geotracking emerges as a mission-critical mean to observe, track and otherwise manage mobile entities from the regular notification of their geographic locations, thanks to positioning devices such as GPS and cell phones. First used in logistic systems, geotracking now strives for large-scale with new applications like the management continent-wide green tax collection on millions of trucks from their passage through thousands virtual toll points. As the current single device, fixed position notification frequency policies can no longer cope with the stringent requirements of such large-scale applications, this paper first examines the challenges and opportunities for self-adaptation in geotracking systems and applications. After identifying the main requirements to that end, it then provides for a first set of decision-making models for identified self-adaptation scenarii, in the context of a set of everyday geotracking business-oriented applications. Finally, it proposes a set of tools and the software architecture currently developed under the French ANR project SALTY to address these issues in the conceptual framework of autonomic computing.