I'd like an email alert where a scheduled task has thrown an exception (for example).
Some way to be certain the system is healthy would be great. It's scraping a lot of different endpoints and we cannot rely on them to behave consistently -- special cases are constantly cropping up.