Error Handling

Bulkhead Pattern

A resilience pattern borrowed from ship design that partitions system resources — thread pools, connection pools, semaphores — into isolated compartments so a failure or slowdown in one service cannot exhaust shared resources and bring down unrelated services. For example, HTTP calls to a slow third-party API get their own bounded thread pool, preventing them from consuming all threads and blocking calls to fast internal services. Hystrix and Resilience4j implement the bulkhead pattern alongside circuit breakers.

관련 프로토콜

가이드에서 언급됨

관련 항목