Using Caffeine Cache in Spring Boot: Algorithms, Configuration, and Code Examples

The article begins by recalling Guava Cache’s features—thread‑safe get/put, expiration, and eviction—and notes its reliance on LRU. It then introduces Caffeine Cache, which builds on Guava’s ideas but adopts the modern W‑TinyLFU algorithm to achieve near‑optimal hit rates.

Algorithm advantages : Caffeine combines LFU and LRU via W‑TinyLFU, using a Count‑Min Sketch to track recent access frequencies with a sliding‑window reset that halves counters after a fixed number of inserts, thus handling both steady‑state and bursty workloads.

Cache filling strategies :

Manual loading – specify a function in cache.get(key, k -> computeValue(k)).

Synchronous loading – provide a CacheLoader when building the cache.

Asynchronous loading – use AsyncLoadingCache with CompletableFuture for non‑blocking loads.

Eviction policies :

Size‑based eviction via .maximumSize() or .maximumWeight().

Time‑based eviction with .expireAfterWrite(), .expireAfterAccess(), or a custom Expiry implementation.

Reference‑based eviction using .weakKeys(), .weakValues(), or .softValues() (mutually exclusive).

Statistics : Enable with .recordStats() and retrieve hit rate, eviction count, and load penalty via cache.stats().

Spring Boot integration :

Include the dependencies:

<dependency>
    <groupId>org.springframework.boot</groupId>
    <artifactId>spring-boot-starter-cache</artifactId>
</dependency>
<dependency>
    <groupId>com.github.ben-manes.caffeine</groupId>
    <artifactId>caffeine</artifactId>
    <version>2.6.2</version>
</dependency>

Enable caching with @EnableCaching and configure Caffeine via properties (e.g.,

spring.cache.caffeine.spec=initialCapacity=50,maximumSize=500,expireAfterWrite=10s

) or Java beans. A sample bean creates a SimpleCacheManager with two caches (userCache, deptCache) using Caffeine.newBuilder() and sets TTL and maximum size.

Define a CacheLoader<Object,Object> bean when using refreshAfterWrite so that the cache can refresh entries after the configured interval.

Use Spring’s cache annotations: @Cacheable(value="userCache", key="#id", sync=true) for read‑through caching. @CachePut(value="userCache", key="#user.id") for updates. @CacheEvict(value="userCache", key="#user.id") for deletions.

Annotations support SpEL for keys, conditions, and synchronization, and can be combined with @Caching or @CacheConfig for class‑level defaults.

The article also warns about incompatibilities (e.g., AsyncLoadingCache cannot use weak/soft references, CacheWriter cannot be used with weak keys), and provides best‑practice tips for choosing eviction strategies based on workload characteristics.