Production

Profiles, Actuator, Resilience

What separates a working local app from a production-ready service. The 20% of config decisions that cover 80% of real scenarios.

Environment

Profiles & Configuration

application.yml for base config, application-{profile}.yml for environment overrides. Set spring.profiles.active via environment variable in production.

Property Source Priority (Highest → Lowest)

1. Command line args (--server.port=9090)
2. Environment variables (SERVER_PORT)
3. application-{profile}.yml
4. application.yml
5. @PropertySource annotations
6. Defaults

@ConfigurationProperties — type-safe binding with validation. Prefer over @Value for groups of related properties. Catches typos at startup, not at runtime.

Secrets

Externalized Config

Never commit secrets — passwords, API keys, DB credentials. Options by environment:

Environment	Strategy
Local dev	`application-local.yml` (gitignored)
CI/CD	Environment variables
Production	Vault / AWS Secrets Manager / K8s Secrets

@ConfigurationProperties with @Validated — catch missing required config at startup, not at runtime. Spring Cloud Config Server is available for centralised config in microservices.

Observability

Actuator

Production-ready endpoints baked in — no extra code. Add spring-boot-starter-actuator and configure exposure.

Endpoints You Care About

Endpoint	What It Shows
`/actuator/health`	UP/DOWN + component health (DB, cache, disk)
`/actuator/info`	Build version, git commit, custom properties
`/actuator/metrics`	JVM, HTTP request counts, custom metrics via Micrometer
`/actuator/env`	All resolved config properties (sensitive — lock down)
`/actuator/loggers`	View and change log levels at runtime without restart
`/actuator/httptrace`	Last 100 HTTP requests (useful for debugging)

Health Endpoint Anatomy

GET /actuator/health

{
  "status": "UP",           ← overall status
  "components": {
    "db": { "status": "UP" },          ← DataSource check
    "redis": { "status": "UP" },       ← Cache check
    "diskSpace": { "status": "UP" }    ← Disk check
  }
}

Security: Never expose /actuator/env or /actuator/shutdown publicly. Use management.endpoints.web.exposure.include=health,info,metrics. Implement custom HealthIndicator for downstream service checks.

Performance

Caching

@EnableCaching on a @Configuration class
@Cacheable("users") — on read methods. Key derived from method params by default.
@CacheEvict — invalidate cache on write. Call on save/delete methods.
@CachePut — update cache entry. Less common.

Default cache: ConcurrentHashMap (in-memory, no TTL, not distributed). Production: Redis (spring-boot-starter-data-redis) — supports TTL and distributed caching.

Gotcha: @Cacheable on a private method or called from within the same class is silently ignored — AOP proxy bypass, same root cause as @Transactional.

Stability

Resilience Patterns

One slow downstream call can exhaust threads and cascade. Resilience4j provides four key patterns:

Pattern	What It Does	When to Use
Circuit Breaker	Stops calling a failing service after threshold	Any external HTTP call
Retry	Retries on transient failures with backoff	Idempotent operations only
Rate Limiter	Limits call rate to downstream service	Third-party APIs with rate limits
Bulkhead	Isolates threads per downstream service	Prevent one slow call from starving others

Circuit Breaker State Machine

Quality

Testing Strategy

Type	Annotation	What It Loads	Use For
Unit	`@ExtendWith(MockitoExtension)`	Nothing — plain Java	Service logic
Web slice	`@WebMvcTest`	Only web layer	Controller, validation
Data slice	`@DataJpaTest`	Only JPA + H2	Repository queries
Integration	`@SpringBootTest`	Full context	End-to-end flows

Key Distinctions

@MockBean registers in Spring context, @Mock is pure Mockito — different lifecycle
Testcontainers — real Postgres/MySQL in Docker. Use @DynamicPropertySource for JDBC URL.
Test slices are faster than @SpringBootTest — use them for the bulk of your tests

LearningTree

Production — Profiles, Actuator, Config, Resilience | Spring Guide