Redis Cache Invalidation and TTL Strategy Guide

You are a senior backend engineer specializing in distributed caching. Design a Redis cache invalidation and TTL strategy for:

Application/service: [app name and purpose]
Stack: [language, framework, database, queue, hosting]
Data to cache: [objects, queries, API responses, sessions, permissions, pricing, etc.]
Read/write patterns: [traffic volume, hot paths, update frequency]
Freshness requirements: [seconds/minutes/hours, eventual consistency tolerance]
Failure tolerance: [what happens if Redis is down or stale]
Current pain: [slow queries, high API cost, rate limits, database load, user-facing latency]
Multi-tenant or auth constraints: [tenant isolation, roles, per-user visibility]
Invalidation events: [create/update/delete, webhook, background job, admin action]
Operational constraints: [memory limits, cluster mode, eviction policy, monitoring tools]

Produce:
1. Cache candidate assessment with what should and should not be cached
2. Key naming convention with tenant/user/version safety
3. TTL matrix by data type with rationale
4. Invalidation triggers and event flow
5. Read-through/write-through/cache-aside recommendation
6. Stampede prevention and stale-while-revalidate approach
7. Redis failure fallback behavior
8. Observability metrics, logs, and alerts
9. Testing plan for stale data, race conditions, and permission leaks
10. Incremental rollout plan with feature flags and rollback steps

Prioritize correctness and debuggability over clever caching tricks.