A first question that a designer should consider regarding adding a fault indicator is:
What are the risks that a critical component fails, and nobody knows about it?
If those instances might issue a problem, then the second question is:
Can we rely on the component MTBF?
MTBF is a statistics. What if the component happens to live in the tail of the distribution? Will the stakeholders tolerate black swans due to such failures?
If the stakeholders may tolerate a latent failure associated with the component, then it might be the case that the component is not essential for the system operation. It may be redundant, and it is a good idea to simplify the system by removing the component, and the associated functions, from the system design. Otherwise, if the stakeholders might be furious about a latent hazard, the design should provide means to detect it and to report about it.
Updated on 19 Jul 2016.