It's a mix of different life-enders. Overall, you have an expected in-service life of the boat (which is usually exceeded). Within that life you tend to have separate drivers for different elements of the boat.
The hull fatigue life is driven by the total number of dives - but it's a complex equation that changes depending on the operational use of the boat. Slow dives to deepish depths will have less of an effect than fast changes of depth across a number of depths.
Then you have the core/fuel life. That tends to be based on assumptions as to time spent at various power levels - with a bit of non-linear reactor chemistry thrown in.
Then you have componentry within the NSRP which can be subject to embrittlement, fatigue or other degradation - some of which can be predicted, some not. Because its within the RC its a beggar to fix, so usually means end of life. Lots of the PWR1 boats suffered these sorts of issues.
Finally, you get into things like cable life. Cable insulation has a finite life - once you need to start recabling a boat, the sheer cost of it is usually a life ender too.
None of the above are particularly interrelated, so its a complex balance across many factors.