fix(caching): bound and prioritise cache warming repository reads

[chainsmt]

Bounds all cache-warming repository reads and prevents startup warming from delaying readiness.

• Cap every warming query with an explicit take sized from a configurable CACHE_WARM_MAX_ENTRIES (ConfigService/env, default 100); replaces the unbounded published-courses find that loaded the whole table into memory.
• Select the warmed set by a real priority signal: published courses ranked by enrollment count (createdAt tie-breaker), users by most-recent activity (lastLoginAt DESC).
• Process the working set in bounded batches (CACHE_WARM_BATCH_SIZE, default 25) with a short delay between batches (CACHE_WARM_BATCH_DELAY_MS, default 50ms) to bound DB/memory pressure.
• Export warming metrics: cache_warm_entries_total, cache_warm_last_entries, and cache_warm_duration_seconds.
• Defer startup warming from a blocking OnModuleInit to OnApplicationBootstrap dispatched via setImmediate (not awaited), so warming runs after the server is accepting traffic and never delays readiness.

Testing

• npx tsc --noEmit — clean
• npm run lint — clean
• npx jest src/caching/cache-warming — 2 suites, 10 tests pass (bounding cap, enrollment ordering, batching/delay, metrics, deferred non-blocking bootstrap). Pre-existing unrelated caching spec failures are unchanged.

Closes #1053

[drips-wave]

@chainsmt Great news! 🎉 Based on an automated assessment of this PR, the linked Wave issue(s) no longer count against your application limits.

You can now already apply to more issues while waiting for a review of this PR. Keep up the great work! 🚀

Learn more about application limits

[RUKAYAT-CODER]

Thank you for contributing to the project.