The scheduled reporting feature creates a recurring task to generate reports on a specified cadence. Typically for recurring tasks, we do not retry if the task fails, however we've found that with reporting jobs (which retry up to 3 times), the first attempt often fails while subsequent attempts succeed due to caching (initial load on first attempt takes a long time but subsequent attempts succeed because ES is warmed up). Given this, we should investigate retrying the scheduled reports as well.

This can be done either at a task manager level by allowing recurring tasks to be retried with a flag on the task type definition, or within the scheduled report task by building the retry into the task logic.