Fix: HTTP client connection pool leak causing agent failures

[Zakaria-Kofiro]

HTTP CLIENT CONNECTION LEAK FIX:

CHANGES:

• Always consume HTTP response entity regardless of status code (fixes CLOSE_WAIT accumulation)
• Add response.close() in finally block to ensure connection release
• Set setConnectionManagerShared(false) to let HttpClient manage pool lifecycle
• Add defensive BR/GZIP decompression with null/empty checks

ROOT CAUSE:
Response bodies weren't consumed for non-2xx status codes, leaving connections stuck in CLOSE_WAIT state. Adding BR support amplified the issue by throwing EOFException on empty bodies.

VALIDATION:
2+ hour A/B test at steady state (~19K concurrent connections)

• Baseline: 1,400+ leaked CLOSE_WAIT, agents crash with too much load
• Fix: CLOSE_WAIT cycles 2-110 (healthy), zero crashes
• CPU/Memory: Fix uses same or less CPU/Memory (~10% less)
• FDs: 50% reduction (22K → 8-12K)

Please make sure these check boxes are checked before submitting

• ** Squashed Commits **
• ** All Tests Passed ** - mvn clean test -P default

** PR review process **

• Requires one +1 from a reviewer
• Repository owners will merge your PR once it is approved.