Debugging KCNA Practice Questions (Page 4)

Questions 16–20

1. 16application · medium

   A team notices that a Kubernetes Deployment is using more memory over time and eventually the Pods are killed by the OOMKiller. The application is stateless and can be restarted. What is the best long-term solution?

   • ARestart the Pods periodically using a CronJob.
   • BSet a memory limit and configure a HorizontalPodAutoscaler based on memory usage.
   • CInvestigate the application code for a memory leak and fix it.
   • DIncrease the memory limit to give the application more headroom.
   Reveal Solution Discussion Select an answer first
2. 17foundation · easy

   What is the primary value of application logs in debugging cloud native applications?

   • AThey provide a chronological record of events and errors that can be searched and correlated
   • BThey provide real-time performance metrics for capacity planning
   • CThey replace the need for monitoring metrics and health checks
   • DThey automatically fix the underlying issue when an error occurs
   Reveal Solution Discussion Select an answer first
4. 18foundation · easy

   Which type of metric would be most useful for detecting a sudden increase in error rates in a cloud native application?

   • ANumber of running replicas
   • BMemory usage of the application container
   • CHTTP response status codes counted over time
   • DCPU usage of the application container
   Reveal Solution Discussion Select an answer first
5. 19application · medium

   A platform team wants to detect when a Kubernetes Deployment is unable to keep up with traffic, causing increased latency. They already have Prometheus scraping metrics from the application. Which metric should they primarily monitor to detect this condition?

   • ACPU usage of the Pods
   • BRequest latency percentiles (e.g., p99)
   • CMemory usage of the Pods
   • DNumber of replicas in the Deployment
   Reveal Solution Discussion Select an answer first
6. 20expert · hard

   A team runs a web application in Kubernetes with three replicas. They notice that one replica is receiving a disproportionate amount of traffic, causing it to be overloaded while the others are idle. The Service is configured with `sessionAffinity: ClientIP`. What is the most likely cause?

   • AThe Pods have different resource limits, causing the scheduler to place them on different nodes with varying capacity.
   • BThe Service's `externalTrafficPolicy` is set to `Local`, which routes traffic only to Pods on the node that received the request.
   • CThe `sessionAffinity` is set to `ClientIP`, which routes all requests from the same client IP to the same Pod, causing uneven load if a few clients generate most of the traffic.
   • DThe Deployment's `replicas` is set to 1, but the Service has multiple endpoints.
   Reveal Solution Discussion Select an answer first