So far, all tasks are scheduled all the same.

We will need to differentiate between those recurring tasks (coming from listers' output), and new higher prioritized oneshot tasks (coming from new origin events).

All those tasks will need to get ingested (no starvation).

For this, we introduce the priority notion in tasks (3 values: high, normal, low).
The scheduling policy will be adapted to trigger a given percentage for all sort of prioritized tasks.

We'll begin with the following percentage (which could be adapted later on):

• 0.5: high
• 0.3: normal
• 0.2: low