circuit-breaker

To stop retries on a broken API endpoint, implement a circuit breaker pattern that transitions to an open state after a specified number of consecutive failures, immediately blocking further requests and conserving system resources. This prevents cascading errors across your application.

The circuit breaker pattern is a fault tolerance mechanism that manages three states: Closed for normal operations, Open to block operations after consecutive failures, and Half-open to probe service availability after a cool-down period. This state management halts operations known to be failing.

Configure failure thresholds and cool-down periods for external API calls by setting the maximum number of consecutive failures allowed before the circuit opens and defining the duration of the cool-down period before transitioning to a half-open probing state.

Yes, the circuit breaker pattern works for batch jobs prone to transient failures. It halts operations after a specified number of consecutive failures, preventing resource exhaustion and stopping cascading errors in batch processing just as effectively as with API calls.

Use a circuit breaker instead of standard retry logic when an external API or operation starts returning persistent errors like 500s. Standard retries risk resource exhaustion, while the circuit breaker opens after a configurable failure threshold to immediately return an error or trigger a fallback.