Play Framework

refactor(test): drop unused JpaLazyPoolTest scaffolding from debug detour

The waitForActiveToSettle helper (commit b2111c7b3) and the parkNanos
sampler (commit 37d620fc7) were added during the Windows CI debug pass
on the assumption that the failure was a baseline-cleanup race or a
watcher sampling-resolution issue. Neither was the real cause — the
failure was the warm-pool /countRo lease being too brief to observe,
fixed by switching that assertion to a response-body probe (67dd859).

With the body probe in place these defensive bits are no-ops:
- All four waitForActiveToSettle calls run against an already-settled
  pool (active was already 0 each time) and return immediately.
- parkNanos(10us) sampled 1500x more often than necessary; /count's
  cold-pool lease takes ms and Thread.sleep(1) catches it everywhere.

Net -23 lines. Local ant integration-test still 31/31 green.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>