Spring Boot Actuator Skill

Overview

• Deliver production-ready observability for Spring Boot services using Actuator endpoints, probes, and Micrometer integration.
• Standardize health, metrics, and diagnostics configuration while delegating deep reference material to references/.
• Support platform requirements for secure operations, SLO reporting, and incident diagnostics.

When to Use

• Trigger: "enable actuator endpoints" – Bootstrap Actuator for a new or existing Spring Boot service.
• Trigger: "secure management port" – Apply Spring Security policies to protect management traffic.
• Trigger: "configure health probes" – Define readiness and liveness groups for orchestrators.
• Trigger: "export metrics to prometheus" – Wire Micrometer registries and tune metric exposure.
• Trigger: "debug actuator startup" – Inspect condition evaluations and startup metrics when endpoints are missing or slow.

Quick Start

<!-- Maven -->
<dependency>
    <groupId>org.springframework.boot</groupId>
    <artifactId>spring-boot-starter-actuator</artifactId>
</dependency>

// Gradle
dependencies {
    implementation "org.springframework.boot:spring-boot-starter-actuator"
}

After adding the dependency, verify endpoints respond:

curl http://localhost:8080/actuator/health
curl http://localhost:8080/actuator/info

Instructions

1. Add Actuator Dependency

Include spring-boot-starter-actuator in your build configuration.

Validate: Restart the service and confirm /actuator/health and /actuator/info respond with 200 OK.

2. Expose Required Endpoints

• Set management.endpoints.web.exposure.include to the precise list or "*" for internal deployments.
• Adjust management.endpoints.web.base-path (e.g., /management) when the default /actuator conflicts with routing.
• Review detailed endpoint semantics in references/endpoint-reference.md.

Validate: curl http://localhost:8080/actuator returns the list of exposed endpoints.

3. Secure Management Traffic

• Apply an isolated SecurityFilterChain using EndpointRequest.toAnyEndpoint() with role-based rules.
• Combine management.server.port with firewall controls or service mesh policies for operator-only access.
• Keep /actuator/health/** publicly accessible only when required; otherwise enforce authentication.

Validate: Unauthenticated requests to protected endpoints return 401 Unauthorized.

4. Configure Health Probes

• Enable management.endpoint.health.probes.enabled=true for /health/liveness and /health/readiness.
• Group indicators via management.endpoint.health.group.* to match platform expectations.
• Implement custom indicators by extending HealthIndicator or ReactiveHealthContributor; sample implementations in references/examples.md#custom-health-indicator.

Validate: /actuator/health/readiness returns UP with all mandatory components before promoting to production.

5. Publish Metrics and Traces

• Activate Micrometer exporters (Prometheus, OTLP, Wavefront, StatsD) via management.metrics.export.*.
• Apply MeterRegistryCustomizer beans to add application, environment, and business tags for observability correlation.
• Surface HTTP request metrics with server.observation.* configuration when using Spring Boot 3.2+.

Validate: Scrape /actuator/prometheus and confirm required meters (http.server.requests, jvm.memory.used) are present.

6. Enable Diagnostics Tooling

• Turn on /actuator/startup (Spring Boot 3.5+) and /actuator/conditions during incident response to inspect auto-configuration decisions.
• Register an HttpExchangeRepository (e.g., InMemoryHttpExchangeRepository) before enabling /actuator/httpexchanges for request auditing.
• Consult references/endpoint-reference.md for endpoint behaviors and limits.

Validate: /actuator/startup and /actuator/conditions return valid JSON payloads.

Examples

Basic – Expose health and info safely

management:
  endpoints:
    web:
      exposure:
        include: "health,info"
  endpoint:
    health:
      show-details: never

Intermediate – Readiness group with custom indicator

@Component
public class PaymentsGatewayHealth implements HealthIndicator {

    private final PaymentsClient client;

    public PaymentsGatewayHealth(PaymentsClient client) {
        this.client = client;
    }

    @Override
    public Health health() {
        boolean reachable = client.ping();
        return reachable ? Health.up().withDetail("latencyMs", client.latency()).build()
                         : Health.down().withDetail("error", "Gateway timeout").build();
    }
}

management:
  endpoint:
    health:
      probes:
        enabled: true
      group:
        readiness:
          include: "readinessState,db,paymentsGateway"
          show-details: always

Advanced – Dedicated management port with Prometheus export

management:
  server:
    port: 9091
    ssl:
      enabled: true
  endpoints:
    web:
      exposure:
        include: "health,info,metrics,prometheus"
      base-path: "/management"
  metrics:
    export:
      prometheus:
        descriptions: true
        step: 30s
  endpoint:
    health:
      show-details: when-authorized
      roles: "ENDPOINT_ADMIN"

@Configuration
public class ActuatorSecurityConfig {

    @Bean
    SecurityFilterChain actuatorChain(HttpSecurity http) throws Exception {
        http.securityMatcher(EndpointRequest.toAnyEndpoint())
            .authorizeHttpRequests(c -> c
                .requestMatchers(EndpointRequest.to("health")).permitAll()
                .anyRequest().hasRole("ENDPOINT_ADMIN"))
            .httpBasic(Customizer.withDefaults());
        return http.build();
    }
}

More end-to-end samples are available in references/examples.md.

Best Practices

• Keep SKILL.md concise and rely on references/ for verbose documentation to conserve context.
• Apply the principle of least privilege: expose only required endpoints and restrict sensitive ones.
• Use immutable configuration via profile-specific YAML to align environments.
• Monitor actuator traffic separately to detect scraping abuse or brute-force attempts.
• Automate regression checks by scripting curl probes in CI/CD pipelines.

Constraints and Warnings

• Avoid exposing /actuator/env, /actuator/configprops, /actuator/logfile, and /actuator/heapdump on public networks.
• Do not ship custom health indicators that block event loop threads or exceed 250 ms unless absolutely necessary.
• Ensure Actuator metrics exporters run on supported Micrometer registries; unsupported exporters require custom registry beans.
• Maintain compatibility with Spring Boot 3.5.x conventions; older versions may lack probes and observation features.
• Never expose actuator endpoints without authentication in production environments.
• Health indicators should not perform expensive operations that could impact application performance.
• Be cautious with /actuator/beans and /actuator/mappings as they reveal internal application structure.

Reference Materials

• Endpoint quick reference
• Implementation examples
• Official documentation extract
• Auditing with Actuator
• Cloud Foundry integration
• Enabling Actuator features
• HTTP exchange recording
• JMX exposure
• Monitoring and metrics
• Logging configuration
• Metrics exporters