Life and Survival dataset

[#1]

I am looking at a component within our product that is on 10000+ products in the field. Should I only be looking at failures? I have the ability to include all the products that have never had this component fail. Is it best to include those in the dataset, or should I only be looking at failures?

[#2]

Answer should depend on the details of your data and objectives.

Here is one scenario:

If your data is like what are in this example: Example of the Life Distribution Platform and the objective is to estimate failure probability like in the example, excluding products that have not failed will severely mislead you.