I’m planning to search multiple different sets of multiple indices at once. I’d also like to use
search_after with point-in-time indices for deep pagination. I’ve got some general questions about how/if PIT works in this scenario.
_pitendpoint with multiple indices works fine, but I’m not sure exactly how it works - is the PIT index coupled to the comma-separated set of indices I pass in my call to
/index-1,index-2/_pit?keep_alive=15mwill open a PIT id usable with any search where I want results from
index-1,index-2)? Also, will the implicit
_shard_doctiebreaking work when creating a multi-index PIT index?
The guidance on the elastic blog here re: having a background process create a PIT for use with all search requests (rather than creating one on each search request) seems to contradict the PIT docs which state the following — I must be misunderstanding one of these statements?
The open point in time request and each subsequent search request can return different id; thus always use the most recently received id for the next search request.