Rationale behind autovacuum_naptime = 1 in high-throughput OLTP

[umutoguz]

Hi,
I have been reviewing the autobase configuration and noticed that autovacuum_naptime is set to 1 (1 second), which is significantly more aggressive than the default 1min.
I am running a high-traffic OLTP PostgreSQL 15 cluster (approx. 2-3k TPS) with daily partitioning on SAN storage. I am trying to understand the specific trade-offs of this setting.
My Question:
Is the primary intention of setting naptime = 1 to reduce the "bloat window" by ensuring the autovacuum launcher wakes up immediately after a worker finishes? Or is it targeted at environments with multiple databases within the same cluster to reduce the check interval per DB?
I am concerned about the potential CPU overhead of the launcher waking up every second versus the benefit of faster reaction times. Would you recommend keeping this at 1s for a heavy write-intensive workload?

Thanks for your insights!