As the military sector digitalizes, there is a broader need for distributed computation and storage solutions deployable at the operational edge, which pose the potential of minimizing the conventional information exchanges by among others, proved enhancements in terms of latency and communication cost. The state of the art refers to multiple applications of edge computing with a vast value on tactical clouds, with may range from enabling optimized transmissions, to federated machine learning and data fusion services. However, very few of them explore the risks they entail in terms of sustainability, were an inappropriate orchestration of the edge capabilities may for example, mistakenly or maliciously deplete the energy sources that feed their supportive infrastructure, or cause electromagnetic alterations exploitable by electronic warfare actuations. In order to facilitate their understanding and contribute to their mitigation, this paper delves into the problematic inherent in Tactical Denial of Sustainability (TDoS) against tactical clouds. The conducted research introduces a framework to prevent TDoS and if required, identify their symptoms and mitigate their damage. The proposal is settled on the Self-Organizing Network (SON) paradigm and its Self-Protection (SP) applications, which benefits of their synergy with the Observe-Orient-Decide-Act (OODA) loop towards dynamically and adaptively triggering proportional reactions.