The present paper discusses techniques for k-out-of-n redundant systems based on condition assessment by diagnostics. Key is to assess the time development of the hazard rate, particularly for wear processes. This enables remaining useful life prognostics. Knowledge of acceleration due to stress enhancement enable prognostics for various scenarios. The paper discusses the framework of diagnostics and hazard rate, the associated predictability of failure as well as the impact of k-out-of-n redundancy. The paper concludes with a Markov analysis of a system with time varying transition rates.