Circuit Breakers & Resilience Patterns

Question 1

In the circuit breaker pattern, what does the CLOSED state mean?

Accepted Answer

The circuit is working normally and requests flow through to the downstream service — CLOSED is the normal operating state. Requests pass through the breaker to the downstream service as usual. The breaker monitors failures and will only trip open if the failure count crosses the threshold.

Question 2

What happens to incoming requests when a circuit breaker is in the OPEN state?

Accepted Answer

They are rejected immediately without attempting the downstream call — In the OPEN state the breaker "fails fast" — it rejects requests in microseconds without ever touching the downstream service. This is the core benefit: callers get an immediate error instead of blocking for a full network timeout.

Question 3

What triggers a transition from CLOSED to OPEN?

Accepted Answer

A configured number of consecutive (or windowed) failures crossing a threshold — The breaker only trips when failures accumulate past a threshold (e.g. 3 in a row, or 50% in a 10-second window). A single transient failure should not trip the breaker — that would make the system too sensitive to momentary blips.

Question 4

What is the purpose of the HALF-OPEN state?

Accepted Answer

To send one or a few test requests to check whether the downstream service has recovered — HALF-OPEN is a cautious probe state. Rather than slamming the downstream with full traffic immediately, it lets through only one (or a small batch of) test requests. If they succeed, the breaker closes; if they fail, it opens again and resets the cooldown.

Question 5

Why is "fail fast" preferable to waiting for a timeout on every request when the downstream is down?

Accepted Answer

Blocking threads while waiting on a timeout exhausts your thread pool, causing your own service to become unavailable — Every blocked thread waiting on a 30-second timeout is a thread not available to serve other requests. If enough threads are blocked, your service runs out of worker threads entirely — a condition called thread-pool exhaustion — and your service also goes down even though the bug is in the dependency.

Question 6

What is "cascading failure" in the context of microservices?

Accepted Answer

When one failing dependency causes its callers to fail, which causes their callers to fail, bringing down the whole system — Cascading failure is the domino effect: Service A depends on B; B goes down; A's threads all block on B; A's threads fill up; A goes down; everything upstream of A now also fails. The circuit breaker breaks the chain by stopping requests to B before A's resources are exhausted.

Question 7

What is the role of the cooldown (reset timeout) in the circuit breaker pattern?

Accepted Answer

It gives the downstream service time to recover before the breaker tries again — The cooldown keeps the breaker OPEN for a configured period before transitioning to HALF-OPEN. This breathing room prevents hammering an already-struggling service with a constant stream of retries, which could make recovery harder rather than easier.

Question 8

How does a circuit breaker differ from a simple retry mechanism?

Accepted Answer

A circuit breaker stops attempting the call entirely for a period, whereas a retry immediately re-attempts, potentially worsening load on the downstream — Retries can make things worse: if the downstream is overloaded, rapid retries add even more load. A circuit breaker complements retries by halting calls for a cooldown period so the downstream can recover, then uses a single test request (HALF-OPEN) instead of a burst of retries.

Question 9

Which pattern is most closely related to the circuit breaker and is often used alongside it to isolate failure domains?

Accepted Answer

The Bulkhead pattern — The Bulkhead pattern (named after ship compartments) isolates thread pools or connection pools per downstream dependency. Combined with circuit breakers, it ensures that exhausting resources for one downstream cannot starve capacity for other dependencies.

Question 10

After the breaker moves to HALF-OPEN and the single test request fails (downstream is still unhealthy), what should happen?

Accepted Answer

The breaker should return to OPEN and restart the cooldown timer — A failed test in HALF-OPEN means the service has not recovered. The breaker returns to OPEN to block traffic and restarts the cooldown, giving the downstream more time before the next probe attempt.

Trip and recover

Test your knowledge

In the circuit breaker pattern, what does the CLOSED state mean?

The short version

Interview Questions

References & Further Reading

Related topics

Love this visualization?