Most existing studies addressing COTS obsolescence issues in Cyber-Physical Systems (CPSs) have strong emphasis on the sustainment phases. We have identified a gap on methods, processes, and tools for effective COTS risk analysis in the early systems acquisition phases. To fill the gap, this study proposes a taxonomy of COTS-related technical debt in order to support early identification, communication, and assessment of obsolescence risks in CPS system engineering life cycles. The taxonomy contributes to the identification of seven key types of COTS technical debt according to systematic signs discoverable during early COTS activities, which may contribute to obsolescence in later phases. These seven types of COTS technical debt include COTS functionality mismatch, performance mismatch, interoperability difficulty, versioning frequency, documentation and support readiness, and limitation on system evolution. It is expected that such notions will help to increase the efficiency of COTS-based CPS development, readiness, and sustainment, through more informed COTS decision-making to avoid expensive and unaffordable obsolescence issues in the envisioned systems sustainment phases.