Fixup the maximum concurrent queries in Druid used for calculation in ClientQuerySegmentWalker

[LakshSingla]

Description

#14808 introduced auto mode for calculating the size to be reserved for inlining the subquery results per query, however, I incorrectly used druid.broker.http.numConnections as the maximum number of queries that the broker can serve concurrently instead of druid.server.http.numThreads. The former refers to the connections per query that the broker can use to contact data nodes, whereas the latter refers to the number of threads in the broker reserved for serving the queries (which represents the correct figure).

This PR corrects that patch.

---

This PR has:

• been self-reviewed.
  • using the concurrency checklist (Remove this item if the PR doesn't have any relation to concurrency.)
• added documentation for new or modified features or behaviors.
• a release note entry in the PR description.
• added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links.
• added or updated version, license, or notice information in licenses.yaml
• added comments explaining the "why" and the intent of the code wherever would not be obvious for an unfamiliar reader.
• added unit tests or modified existing tests to cover new code paths, ensuring the threshold for code coverage is met.
• added integration tests.
• been tested in a test Druid cluster.

[gianm]

Everything passed except for an IT that has been problematic lately in other PRs as well. So I'll merge this. Thanks @LakshSingla!