- How long without cooling does it take for the samples to be destroyed?

- When it does go down, how long does it take you to catch it?

- Are you doing active "chaos monkey"-style testing such as intentionally turning off random systems and ensuring your backup procedures detect and solve the problem before the samples are destroyed?

- Given the current deterioration rates in your industry and your facility under normal operation, what is the probability of a sample surviving 10 years? 20? 30? What about after accounting for the possibility of equipment/system failures?