You have it backwards. The evidence for a cluster is the cluster now in need of maintenance, and it's the clustering by maintenance cycles and how they occurred which matters, not start date.

Start date is by no means the only thing affecting later clustering of maintenance need. That's also easily achieved by for example deferring maintenance too long, or structuring of past maintenance cycles.

Understanding how to avoid maintenance cycles syncing up is critical for any risk management of large systems.

This is entirely separate from whether or not one supports nuclear, and an issue relevant for any type of generation. Or indeed managing any kind of system that includes any physical plants or machinery at all.

How that affects whether you see a given type of plant as viable is an entirely separate issue.