chore: add postman collection for redis reconnect verification

A standalone collection + environment under postman/ that covers every
endpoint touching Redis, so the auto-reconnect / command-timeout /
readiness probe behaviour can be verified end-to-end:

- Health probes: /hello (Liveness), /health (Readiness) with assertions
  on 200/503 and Cache-Control.
- Direct Redis usage: POST /auth/@login flows that exercise the
  loginLock:* hash (read, write, expire, lock trigger).
- Indirect Redis usage via cache-manager: GET /users/:id and
  /namespaces/:key cache miss/hit/invalidate sequences.

Collection description includes a step-by-step recipe for the actual
outage drill: stop Redis, watch /hello stay 200, /health flip to 503
within ~1.5s, business commands fail fast (~2s) instead of hanging,
then restart Redis and watch /health auto-recover.

Co-authored-by: Cursor <cursoragent@cursor.com>

Author: zzswang