Check Current System Pressure Before Allocating a New Thread In Concurrent Search [LUCENE-9075]

When executing slices for a concurrent query, we should check the current status of the system before allocating since that can contribute to the pressure on the system and further degrade the situation

---
Migrated from [LUCENE-9075](https://issues.apache.org/jira/browse/LUCENE-9075) by Atri Sharma (@atris)
Linked issues:
- #10108

Start by locating the concurrent query code that executes slices and allocates new threads. Review the linked issue #10108 for related context before tracing the allocation path. Done means the system's current pressure is checked before each new thread is allocated, with behavior that avoids worsening pressure during concurrent search.

Written by the indexing model from the issue text.