Circuit Breaker Pattern Simulator

Mental Model

The Circuit Breaker pattern is a critical fault-tolerance pattern used in distributed systems to prevent cascading failures. It wraps network requests to downstream dependencies, failing fast during service degradations to protect upstream worker threads and database pools, and automatically probing for recovery.

The core mental model mirrors an electrical circuit:

• Closed: Normal operation. Requests pass to the downstream service. If failure rates or slow call rates cross thresholds, the breaker trips to Open.
• Open: Fail-fast phase. Requests are immediately blocked and redirected to a Fallback Handler to return cached or default data, bypassing the degraded service.
• Half-Open: Probing phase. After a timeout duration, a limited number of requests are allowed to pass as probes. If they succeed, the breaker returns to Closed. If any fail, it trips back to Open.

For preconfigured stress tests, you can also select the scenario presets directly in the simulator.

Components and Why They Exist

Client App

The client application represents the upstream caller initiating downstream service requests. It generates request traffic at a configurable rate (RPS) and tracks end-to-end metrics like success rate and average latency. Request particles originate at the Client App card and record either live service completions or fast fallback responses on return.

Circuit Breakers

The circuit breakers act as protective proxies wrapping downstream service calls to isolate faults. Each breaker maintains sliding windows of outcomes (failures or slow calls), evaluates trip thresholds, manages state transitions, and fails fast when open. The ring buffer on each breaker card colors recent calls green, red, or yellow while the state badge flips between CLOSED, OPEN, and HALF-OPEN.

Downstream Services

The downstream services are independent microservice dependencies queried by the client through the breaker. They process incoming requests, introduce latency spikes or error rates, or experience complete outages. Error rate badges, latency multipliers, and online/offline status indicators on each service node show the exact fault condition the upstream breaker is reacting to.

Fallback Handler

The fallback handler serves degraded alternative responses when a downstream dependency is unreachable or slow. It intercepts rejected or failed calls and immediately returns a cached or static fallback payload. When a breaker trips open, purple short-circuited particles divert through the Fallback Handler card and return to the Client App without waiting on the failing service.

Edge Cases and Misconceptions

• Sliding Window Sensitivity: Short sliding windows might trip too quickly on transient spikes, while very long windows delay failure detection. Compare Count-based vs Time-based windows under varying traffic rates.
• Slow Calls and Resource Exhaustion: A service that is slow but not failing can tie up threads and cause cascading failures. The slow call rate threshold trips the circuit before threads exhaust.
• Fallback Health: Fallbacks should be simple and fast. If the fallback handler becomes slow, the client still suffers.
• Probing (Half-Open) Under Load: If the probe traffic allowed in the Half-Open state is too high, it can immediately overload a recovering service, driving the circuit back to the Open state.