DevOps Site Reliability Engineering 2 — Questions and Answers
Question 1: What is a circuit breaker pattern?
- A design pattern that stops sending requests to a failing service to allow it to recover, instead of overwhelming it (Correct answer)
- A Kubernetes network policy
- A deployment rollback mechanism
- A load balancing algorithm
Correct answer: A design pattern that stops sending requests to a failing service to allow it to recover, instead of overwhelming it
The circuit breaker opens when a service exceeds a failure threshold, redirecting or failing fast for requests instead of waiting for timeouts, protecting the caller and giving the failing service recovery time.
Question 2: What is 'observability-driven development'?
- Building systems with instrumentation and telemetry from the start so they are intrinsically observable (Correct answer)
- Writing unit tests for monitoring code
- Using observability tools to write code
- Replacing code reviews with log analysis
Correct answer: Building systems with instrumentation and telemetry from the start so they are intrinsically observable
Observability-driven development treats telemetry as a first-class concern, ensuring every feature ships with the metrics, logs, and traces needed to understand its behavior in production.
Question 3: What is 'progressive delivery'?
- Releasing features incrementally to subsets of users using feature flags, canary releases, or A/B testing (Correct answer)
- Delivering features sorted by priority
- A Scrum sprint delivery model
- Deploying code only during business hours
Correct answer: Releasing features incrementally to subsets of users using feature flags, canary releases, or A/B testing
Progressive delivery decouples deployment from release, using techniques like feature flags and canary releases to control who sees new features and roll back instantly if needed.
Question 4: What is a load balancer health check?
- A periodic test sent to backend instances to determine if they are healthy and should receive traffic (Correct answer)
- A network bandwidth monitoring tool
- A CPU utilization alert
- A Kubernetes liveness probe
Correct answer: A periodic test sent to backend instances to determine if they are healthy and should receive traffic
Load balancers use health checks (HTTP, TCP, or custom probes) to automatically remove unhealthy instances from the rotation and restore them once they pass checks again.
Question 5: What is backpressure in distributed systems?
- A mechanism where a downstream service signals upstream services to slow down when it's overwhelmed (Correct answer)
- Network latency between microservices
- A database replication lag
- A Kubernetes resource quota
Correct answer: A mechanism where a downstream service signals upstream services to slow down when it's overwhelmed
Backpressure prevents cascading overload by letting overwhelmed services reject or slow requests, signaling upstream callers to reduce their send rate.
Question 6: What is a 'golden signal' in SRE monitoring?
- One of four key metrics: latency, traffic, errors, and saturation (Correct answer)
- A high-priority production alert
- A 100% uptime milestone
- A Grafana dashboard theme
Correct answer: One of four key metrics: latency, traffic, errors, and saturation
The four golden signals (from Google's SRE book) — latency, traffic, errors, and saturation — are the most critical metrics to monitor for any user-facing service.
DevOps Site Reliability Engineering 2
What is a circuit breaker pattern?