Built‑in In‑Memory Cache vs. Spring Cache for Thymeleaf Fragment Caching: Which Is Better for Stale Data Prevention?

Goal

To determine the most reliable strategy for keeping Thymeleaf fragment data current in a multi‑node application.

Constraints

The built‑in cache is single‑threaded and requires manual clearing; it offers no TTL or automatic invalidation. Spring Cache, while configurable with Redis or Hazelcast, introduces extra wiring and potential latency.

Uncertainty

It is unclear which approach provides acceptable freshness without sacrificing performance or scalability in a clustered environment.

Questions

1. In what deployment scenarios does the built‑in cache lead to stale fragments that cannot be resolved by simple restarts?
2. How do eviction policies (TTL, LRU) in Spring Cache impact rendering latency compared to the static built‑in cache?
3. Would a hybrid pattern—using Spring Cache for long‑lived fragments and the built‑in cache for short‑lived ones—offer a balanced solution?