I've gotten in the habit of throwing an exception that results in an error in the UI if the problem is caused by data or something else the user can fix, raisin

I've gotten in the habit of throwing an exception that results in an error in the UI if the problem is caused by data or something else the user can fix, raising an alert to be handled by the team if there's an issue that needs to be fixed by the team to prevent an outage, and sending daily e-mail reports for things the team wants to monitor (usually temporarily). It's more about the categorization and impact than the implementation. Feel free to send me an e-mail if you have questions.