# Change: Add Backend Reliability Improvements

## Why
Several backend reliability issues need addressing: database connection pool may be undersized for production load, Redis publish failures have no fallback mechanism, tasks with active blockers can be deleted without warning, and NAS storage configuration needs validation.

## What Changes
- Optimize database connection pool configuration
- Add local queue fallback for Redis publish failures
- Add pre-deletion check for active blockers on tasks
- Validate NAS storage path configuration on startup

## Impact
- Affected specs: document-management
- Affected code:
  - `backend/app/core/database.py` - Connection pool config
  - `backend/app/services/notification_service.py` - Redis fallback
  - `backend/app/api/tasks/router.py` - Blocker check
  - `backend/app/services/file_storage_service.py` - NAS validation