SRE Capacity Planning & Scaling 5 — Questions and Answers
Question 1: What is 'graceful degradation' in the context of capacity planning?
- Reducing feature functionality under overload rather than failing completely (Correct answer)
- Gradually decommissioning old infrastructure
- Scaling down resources during off-peak hours
- Handling database failovers automatically
Correct answer: Reducing feature functionality under overload rather than failing completely
Graceful degradation means intentionally shedding non-critical functionality under excess load to preserve core service availability and user experience.
Question 2: An SRE team uses the 'four golden signals' for capacity-related alerting. Which signal most directly indicates imminent capacity exhaustion?
- Errors
- Saturation (Correct answer)
- Latency
- Traffic
Correct answer: Saturation
Saturation measures how full a resource is (CPU, memory, disk, network), making it the most direct indicator that capacity is being exhausted.
Question 3: Which technique allows a team to test system behavior under 2× production load without generating real user traffic?
- Shadow traffic (dark launch) mirroring (Correct answer)
- Feature flag rollout to 50% of users
- Increasing CDN TTLs
- Enabling database read replicas
Correct answer: Shadow traffic (dark launch) mirroring
Shadow traffic mirrors real production requests to a parallel environment, allowing load testing at scale with real patterns without impacting users.
Question 4: In vertical scaling, what is the primary limitation compared to horizontal scaling?
- Vertical scaling requires load balancers
- There is a hardware ceiling beyond which a single machine cannot be scaled further (Correct answer)
- Vertical scaling always costs more than horizontal scaling
- Vertical scaling requires application code changes
Correct answer: There is a hardware ceiling beyond which a single machine cannot be scaled further
Vertical scaling is bounded by the largest available hardware instance, whereas horizontal scaling can theoretically add unlimited nodes.
Question 5: What is 'traffic shaping' used for in capacity management?
- Routing traffic geographically for latency reduction
- Controlling the rate and distribution of incoming requests to protect backend capacity (Correct answer)
- Compressing network packets to reduce bandwidth usage
- Prioritizing HTTPS traffic over HTTP
Correct answer: Controlling the rate and distribution of incoming requests to protect backend capacity
Traffic shaping uses rate limiting, queuing, and throttling to control request flow so backends are not overwhelmed beyond their capacity.
Question 6: A team's capacity plan assumes 20% year-over-year traffic growth, but actual growth is 60%. What SRE process failure does this indicate?
- Insufficient monitoring coverage
- Poor demand forecasting and failure to review forecasts against actuals (Correct answer)
- Inadequate on-call staffing
- Missing circuit breakers in the application
Correct answer: Poor demand forecasting and failure to review forecasts against actuals
A large gap between forecast and actual growth indicates the demand forecasting process lacked accuracy and regular calibration against real traffic trends.
Question 7: Which capacity planning concept describes the point where adding more resources produces diminishing or negative returns due to coordination overhead?
- Saturation point
- Coherency penalty
- Universal Scalability Law (USL) retrograde region (Correct answer)
- Amdahl ceiling
Correct answer: Universal Scalability Law (USL) retrograde region
The Universal Scalability Law models throughput vs. concurrency and predicts a retrograde region where coordination costs cause throughput to decrease with more resources.
SRE Capacity Planning & Scaling 5
What is 'graceful degradation' in the context of capacity planning?