SCAN Search type behavior explanation


I'm working right now on a use-case where I need to index all my data.
I read on elasticsearch forums that SCAN search type will be the best
approach to do that.

I have 2 questions regarding the SCAN/scroll:

  1. I know each scroll has a time to live, I was wondering if after this
    time is elapsed and I scroll another page if the first scroll segments are
    removed if the documents from the first scan got updated.
  2. Another requirement we have is to support sending multiple SCANs at
    the same time, is there any concern running 10-100 SCANs at the same time?
    any memory constraints I need to be aware besides the index size which
    might grow?

Thanks for the help!