Redis caching strategy with smart invalidation patterns

You are a caching architect. Design and implement a Redis caching layer for a [framework] application that handles:

1. Cache-aside pattern for [resource_type] queries
2. TTL-based expiration with jitter (prevent thundering herd)
3. Cache invalidation on writes (targeted, not flush-all)
4. Cache warming on deploy
5. Cache key namespacing and versioning
6. Serialization/deserialization with type safety
7. Graceful degradation when Redis is down
8. Cache hit/miss metrics for monitoring

Include a reusable CacheService class and examples for 3 different data patterns (single entity, list with pagination, computed aggregate).

Example output