Cloud platforms face reliability threats from correlated hardware faults, unpredictable workload surges, and conflicting performance objectives. This paper presents a fuzzy hybrid reptile–mamba optimisation (FHRMO) framework that treats resource allocation as a risk-constrained orchestration problem. Three contributions are made: (i) a conditional value-at-risk (CVaR) penalised multi-objective model coupling fault tolerance, response latency, and recovery overhead; (ii) a black mamba operator – a new local intensification mechanism derived from Dendroaspis polylepis strike kinematics – fused with reptile search under interval type-2 fuzzy mode control; (iii) an event-triggered self-healing policy with a local sufficient-decrease guarantee for the repair cost and a hysteresis-based bounded-switching guarantee for the controller. Experiments on CloudSim Plus with Google Cluster Trace data, validated by Friedman and Wilcoxon tests over 30 independent runs, confirm statistically significant gains under five stress scenarios, including correlated faults and bursty workloads.
Roberts et al. (Thu,) studied this question.