"1 second" vs NULL for continuous aggregate end_offset?

I want to have the most up-to-date data in my continuous aggregate. When I enable real-time, I cannot get the performance I want (especially in the case of multiple continuous aggregates on top of each other). For this reason, I want to run the policies one after another as often as possible.

I see that it is not recommended for end_offset to be NULL. What if I set it to ‘1 second’? Are the intervals specified here determined at the beginning of the job or are they recalculated for each row?

Also my hypertable has 1 hour chunk time interval and my policy config is like below:

SELECT add_continuous_aggregate_policy(
		'record_bucket_1min',
		start_offset => INTERVAL '5 minutes',
		end_offset => interval '1 second',
		schedule_interval => INTERVAL '1 second'
	);```