Skip to main content

Health indicators

When spring-boot-actuator is on the classpath, jEAP Messaging automatically registers a health indicator.

jeapKafka

The jeapKafka indicator checks broker connectivity for each configured Kafka cluster by calling describeCluster() via the Kafka Admin API. It reports clusterId and nodeCount per cluster and is available at /actuator/health/jeapKafka.

Probe integration

Probe integration is opt-in: jeapKafka can be added to the liveness and/or readiness probe groups.

management.endpoint.health.probes.enabled=true
management.endpoint.health.group.liveness.include=livenessState,jeapKafka
management.endpoint.health.group.readiness.include=readinessState,jeapKafka

The same down-after grace period applies to both.

Recovery grace period

The indicator reports UP with a recovering: true detail for a configurable grace period after a failure is first detected, and only flips to DOWN after it elapses — giving Kafka's built-in reconnection time to recover.

The total time before a probe sees DOWN = down-after + (failure threshold × check interval). With the default down-after=35s:

PlatformProbe settingsTime before DOWN
KubernetesfailureThreshold=5, periodSeconds=5≈ 60s
AWS ECSunhealthy_threshold=5, health_check_interval=15s≈ 110s

Properties

PropertyDefaultDescription
management.health.jeap-kafka.enabledtrueEnable/disable the indicator
management.health.jeap-kafka.response-timeout1sTimeout per broker check
management.health.jeap-kafka.down-after35sGrace period before reporting DOWN
management.health.kafka.enabledfalseSpring Boot's built-in single-cluster indicator, disabled by default
spring.kafka.listener.auth-exception-retry-interval10sRetry interval for auth failures in listener containers