Hibernate Cache

Hibernate cache is a mechanism used to reduce database access by storing data in memory. It improves performance by avoiding repeated database queries for the same data.

Hibernate provides multiple cache levels, each serving a different purpose.

Important:
Cache is a performance optimization, not a source of truth.

Cache Levels in Hibernate

Hibernate supports three cache levels:

First-Level Cache
Second-Level Cache
Query Cache

1️⃣ First-Level Cache (Default)

What It Is

Also called Persistence Context
Exists per Session / EntityManager

You cannot disable it.

How It Works

Customer c1 = entityManager.find(Customer.class, 1L);
Customer c2 = entityManager.find(Customer.class, 1L);

Only one SQL query is executed.

Why?

First-level cache stores the entity
Same entity ID → same object instance

All JPA repository methods automatically use the First-Level Cache.
Ex: customerRepository.findById(1L);

Characteristics

Scope: Transaction / Session
Lifetime: Until transaction ends
Stores: Managed entities only
Acts as identity map

Rule:
Within one transaction, Hibernate never hits DB twice for same entity ID.

2️⃣ Second-Level Cache

What It Is

Shared across sessions
Application-level cache
Requires explicit configuration

Used for:

Read-heavy data
Reference data
Rarely changing data

Example Use Cases

Branch details
Product configurations
Country / currency lists
❌ Not for transactional data.

How It Works

Customer c1 = session1.find(Customer.class, 1L); // DB hit
Customer c2 = session2.find(Customer.class, 1L); // Cache hit

Enable Second-Level Cache

Requires cache provider:

Ehcache
Caffeine
Hazelcast

Please refer implementation of Caffeine : Cache Notes

3️⃣ Query Cache (RARELY USED)

What It Is

Caches query results
Depends on second-level cache
Disabled by default

query.setCacheable(true);

Why Query Cache Is Rare

Hard to keep consistent
Sensitive to data changes
Risk of stale results

Most teams avoid it.