For our on-prem installation, we enabled the elastic search integration because we hoped search results would be better, but apparently they aren’t.
Amongst other types of content, we collect Java stack traces in issues. These contain many qualified names of the shape “a.b.C” (words delimited by dot, e.g., “java.lang.IllegalStateException”).
I had expected that I would find tickets with such content by querying for a segment of the qualified name, e.g., “IllegalStateException” but no ticket with such content was found. Only the full name in the query would reveal that content “java.lang.IllegalStateException”.
Since I cannot believe that this is intentional behavior, I wanted to ask if my expectation is wrong.