BreakerPattern

BreakerPattern is a resilience design pattern used in distributed systems to prevent cascading failures when a downstream dependency becomes unreliable. It monitors the outcome of requests to that dependency and, if failures surpass a configured threshold, temporarily blocks further calls and returns a fast failure or a fallback. The aim is to isolate faults, allow the failing service to recover, and maintain overall system responsiveness.

Practically, it is realized as a simple state machine with three states: closed, open, and half-open. In the closed state, calls pass through and results are tracked. If the error rate or failure count exceeds the threshold, the breaker switches to open, short-circuiting subsequent requests for a cooldown period. After the cooldown, it enters half-open to test a limited number of calls. A successful test returns to closed; a failure returns to open. Common configuration includes the failure threshold, cooldown duration, and the size of the half-open probe. Observability of error rate, latency, and throughput guides tuning, and fallbacks are typically provided for open or half-open states.

BreakerPattern is widely used for remote services, databases, and third-party APIs, especially in microservice architectures. Benefits include reduced risk of cascading failures and faster recovery from upstream faults. Drawbacks include added complexity and the potential for latency during half-open testing or mis-tuning, which can cause oscillations. Variants exist that use rolling windows or probabilistic checks, and most implementations combine breakers with timeouts, retries, and fallback strategies to form a resilient stack.