Quiz: Chapter 07 (Resource Management & QoS)
Questions
Why are requests/limits mandatory for production workloads?
What QoS class do pods usually get when requests and limits are both set but not equal?
What Kubernetes object enforces namespace-wide total resource caps?
What Kubernetes object provides default/min/max resource values per container?
Which statement is correct?
- A) BestEffort pods are safest for critical APIs.
- B) ResourceQuota helps prevent one namespace from exhausting cluster capacity.
- C) OOMKilled means the container exceeded CPU limit.
Why include ephemeral-storage requests/limits?
Preferred response to repeated OOMKilled is:
- A) remove limits
- B) inspect memory profile, adjust requests/limits, verify with metrics
- C) disable probes
What is the risk of increasing quotas without capacity review?
Which QoS class is obtained when CPU+memory requests equal limits for every container?
Complete the guardrail:
- A) scale first, investigate later
- B) evidence first (metrics/events), then resource tuning
- C) disable quotas in non-prod by default
Answer Key (Short)
- Predictability, fair scheduling, and controlled blast radius under pressure.
- Burstable.
ResourceQuota.LimitRange.- B
- To control disk pressure and avoid node instability from runaway writable layers/tmp.
- B
- Capacity oversubscription and cross-namespace instability.
- Guaranteed.
- B