See how retries can make a failing dependency even worse — and how a circuit breaker changes the pressure.
What problem are we solving?
A downstream service is failing. Your service must stop turning a dependency failure into a cascading failure.
Live simulation
updates every 1s
CLOSED
Caller→Breaker→Dependency
Controls
20%
5 failures
2x
Live metrics
Effective pressure60%
Breaker stateCLOSED
Retry amplification3×
What just happened?
Retries can multiply traffic toward a dependency that is already unhealthy. An open breaker cuts that feedback loop and lets the caller fail fast or degrade gracefully.
Try this
Dependency outage
Raise dependency failure above 80%, then add retries. Notice how pressure rises while the breaker protects the dependency.
What should happen to user experience while the breaker is open?
Key takeaway
Retries help transient failures. Circuit breakers protect you when the failure is persistent.
