This is what a validation study is for. Pick a number for your run time, do lots of testing to ensure the ice meets specifications.
Once that is done and you have determined a level of risk you're comfortable with, you run normally and do some lower frequency verification testing to make sure the product is staying within the expected/allowed range.
Is the ice consumable or food contact? What is the water source? With reverse osmosis water and consumption a week seems like a reasonable starting point for validation.
Edited by G M, 12 January 2023 - 05:04 PM.