Continue-as-new Vs Child workflow

Hi,
We have a parent workflow that basically performs a long running iterative process. Something like lets say fetch 10k ids from database, and process each id.
In order to prevent event history growing beyond the limit, we thought of 2 approaches

  1. Have batches of 1k ids, for each id start a child workflow and then continue as new when this batch is processed.
  2. Have batches of 1k ids, for each id start a child workflow, and then start the next batch as a child workflow
    In both cases the event history will not grow by a lot so whats the best strategy.

Both would work. I prefer (1) as it works for unlimited dataset size.

Thanks Maxim!