Caching in a microservices setup isn't as simple as just dropping in Redis.
Here’s a quick matrix we used recently:
Use Case | Cache Type | Notes |
Idempotent lookup | Read-through | Use TTL |
Aggregated data | Write-behind | Async flush |
Reference data | Local/Coherent | Watch staleness |
Mistakes we made:
- Cached 404s for too long
- Missed cache invalidation across services
Tools we used: Redis + Caffeine combo.