What is the primary problem a circuit breaker solves?

Circuit breakers prevent cascading failures by fast-failing instead of waiting for timeouts, which exhausts thread pools and causes system-wide outages.

Without a circuit breaker, what happens when an external API goes down?

Without a circuit breaker, each request waits for a timeout (e.g., 30 seconds), consuming a thread. Thousands of waiting threads exhaust the pool and memory, causing the entire application to fail.

What does 'fast-failing' mean in the context of circuit breakers?

Fast-failing means returning an error response immediately without attempting to call the broken downstream service, freeing up threads and resources instantly.

What happens during the CLOSED state of a circuit breaker?

CLOSED is the normal healthy state. Requests pass through to the downstream service, and the circuit breaker tracks failures. If failures exceed a threshold, it transitions to OPEN.

What is the purpose of the HALF-OPEN state?

HALF-OPEN is a recovery test state. The circuit breaker allows a limited number of probe requests through. If they succeed, it returns to CLOSED; if they fail, it returns to OPEN.

What triggers the transition from OPEN to HALF-OPEN?

After the circuit breaker is OPEN, it waits for a configured recovery timeout (e.g., 30 seconds). When that time expires, it automatically transitions to HALF-OPEN to attempt recovery.

Which state transition indicates a successful recovery of the downstream service?

A transition from HALF-OPEN to CLOSED indicates the probe requests succeeded, meaning the downstream service has recovered and normal traffic can resume.

What is the failure threshold in a circuit breaker?

The failure threshold (e.g., '5 failures in 10 seconds') defines when the circuit breaker should trip. Once exceeded, it transitions from CLOSED to OPEN.

What is the recovery timeout in a circuit breaker?

The recovery timeout (e.g., 30 seconds) is how long the circuit breaker waits while OPEN before transitioning to HALF-OPEN to test if the service has recovered.

If your failure threshold is too sensitive (e.g., 2 failures in 10 seconds), what problem occurs?

A too-sensitive threshold causes the breaker to trip on normal blips and transient errors, falsely blocking traffic to a service that is actually healthy.

If your recovery timeout is too long (e.g., 5 minutes), what impact does this have?

A long recovery timeout means the circuit stays OPEN longer before attempting recovery in HALF-OPEN, unnecessarily prolonging unavailability if the service recovers quickly.

Why is fast-failing better than waiting for a timeout?

Fast-failing returns an error immediately instead of tying up a thread for 30+ seconds. This prevents thread pool exhaustion, memory issues, and cascading failures.

What is the relationship between thread exhaustion and cascading failures?

When threads pile up waiting for timeouts, memory depletes and the application becomes unresponsive. This causes it to fail and propagate failures to services depending on it.

How does a circuit breaker improve user experience when a dependency is down?

Instead of users waiting 30+ seconds for a timeout error, the circuit breaker returns an error instantly, allowing the user to see the issue and the app to handle other requests.

What is the main limitation of having a circuit breaker with only local, in-memory state?

In a distributed system with multiple pods, each pod's independent circuit breaker has its own failure count. Some pods may not yet see enough failures to trip, so they keep calling a broken service.

How does using Redis solve the distributed circuit breaker state problem?

Redis acts as a shared store for circuit breaker state. All pods check and update the same failure counter and state in Redis, ensuring consistent decisions across the fleet.

In a Redis-backed circuit breaker, what does TTL (time-to-live) on failure counters accomplish?

Redis keys with TTL expire automatically. This means failure counts reset after the configured window, preventing old transient failures from keeping the breaker OPEN indefinitely.

When multiple pods simultaneously check and increment the failure counter in Redis, what issue could occur without atomic operations?

Without atomic operations (like Redis INCR), concurrent increments from multiple pods could be lost due to read-modify-write race conditions.

In a sports stats platform, when the external Sports API goes down during a World Cup final, what should happen?

When the external API fails, a circuit breaker prevents hammering it and enables fallback to cached scores, maintaining partial availability during the outage.

How should the circuit breaker be configured for a sports stats API that must be highly available during major events?

Major events demand high availability. A lenient threshold tolerates transient blips, and caching fallback ensures users still see content even if live updates are unavailable.

Why is it beneficial to use separate circuit breakers for each notification provider (FCM, SendGrid, Twilio)?

Independent circuit breakers per provider mean if SendGrid is down, push notifications via FCM and SMS via Twilio can still be sent, maximizing delivery success.

In a notification system, what should happen when a circuit breaker is OPEN for the email provider?

When the email circuit is OPEN, the system should send notifications via other available channels (push, SMS) immediately and queue email for later retry, maximizing delivery.

What is a state machine in the context of circuit breakers?

A state machine defines exactly which states exist and what triggers each transition. Circuit breakers use this pattern to ensure predictable, well-defined behavior.

You set a failure threshold of 10 failures in 60 seconds. What does this mean?

The time window is crucial. A 60-second window means only failures in the last 60 seconds count toward the threshold; failures older than 60 seconds are ignored.

What metric would indicate that a circuit breaker is effectively preventing cascading failures?

A working circuit breaker allows the app to free threads and memory by fast-failing. You'd see threads return to the pool and memory stabilize despite downstream failures.

Circuit Breakers — System Design Concept

🍪 TechCookies

Free

📚Circuit BreakersFree

9 sections

~27 min total

25 quick quizzes

3 SD challenges linked

0 of 9 done·~24 min left

Concepts›Circuit Breakers›What is Circuit Breakers?

0 / 9

9 sections~27 min

What is Circuit Breakers?

Prevent cascading failures by stopping requests to failing dependencies and fast-failing instead.

Read

~3 min

⋯

What Problem Does a Circuit Breaker Solve?

Prevents cascading failures by blocking requests to broken services instead of exhausting threads waiting for timeouts.

ReadQuiz

~3 min

⋯

The Three States: CLOSED → OPEN → HALF-OPEN

Circuit breaker transitions between healthy (CLOSED), tripped (OPEN), and recovery (HALF-OPEN) states based on failures and timeouts.

ReadQuizCode

~3 min

⋯

Failure Threshold and Recovery Timeout Configuration

Two configuration parameters control when the breaker trips and when it attempts recovery, balancing sensitivity and responsiveness.

ReadQuizCode

~3 min

⋯

Fast-Fail: Why Speed Matters When a Dependency is Down

Returning errors immediately instead of waiting for timeouts prevents thread pool exhaustion and cascading system failures.

ReadQuizCode

~3 min

⋯

Shared Circuit Breaker State Across Multiple Pods (Redis)

Distributed systems require circuit breaker state to be shared across pods via Redis to maintain consistent behavior and prevent hammering failing services.

ReadQuizCode

~3 min

⋯

Relevant System Design Question: Sports Stats API Resilience

Design a sports score platform that handles external API failures using circuit breakers, caching, and graceful degradation strategies.

ReadQuizCode

~3 min

⋯

Relevant System Design Question: Notification System

Design a notification system with independent circuit breakers per provider (FCM, SendGrid, Twilio) to isolate failures and maintain partial availability.

ReadQuiz

~3 min

⋯

Practice test

25 questions

~9 min

Section 1 of 9Read

What is Circuit Breakers?

Prevent cascading failures by stopping requests to failing dependencies and fast-failing instead.

~3 min read

Circuit breakers are one of the most important resilience patterns in modern distributed systems. Just like an electrical circuit breaker in your home stops the flow of electricity when something goes wrong (preventing fires), a software circuit breaker stops the flow of requests to a failing service — protecting your system from cascading failures and giving the broken service time to recover. If you have ever seen a website show "Service temporarily unavailable" instead of hanging forever, there is a good chance a circuit breaker was doing its job.

Mark this section complete.

Notes

🔍

Loading…