It's possible if rare, for the order processor to reset the database and drop all tasks while robot allocator is still running. If the robot allocator had jobs associated with tasks that are now dropped, it may still try to complete that task even if it doesn't exist anymore.
While this will not crash the system, it does cause the robots to move an item to the station that no longer has that task/order, so it's wasted time.
Instead, one option is on job updates, to check if the task still exists, and if not to drop the item/job/assignment for that robot.
This might add too many extra redis queries though, so perhaps only do this check on item pickup/drop.
It's possible if rare, for the order processor to reset the database and drop all tasks while robot allocator is still running. If the robot allocator had jobs associated with tasks that are now dropped, it may still try to complete that task even if it doesn't exist anymore.
While this will not crash the system, it does cause the robots to move an item to the station that no longer has that task/order, so it's wasted time.
Instead, one option is on job updates, to check if the task still exists, and if not to drop the item/job/assignment for that robot.
This might add too many extra redis queries though, so perhaps only do this check on item pickup/drop.