6.2 Configuring Health Monitoring and Alerting PROFESSIONAL-SECURITY-OPERATIONS-ENGINEER Practice Questions (Page 1)

Questions 1–5

1. 1application · medium

   You are designing a health monitoring strategy for a new microservices-based application. You need to ensure that you can detect when a service is unhealthy, but you also want to minimize the number of alerts that require manual investigation. Which set of metrics should you prioritize?

   • ADisk I/O and network throughput for each service.
   • BRequest latency and error rates for each service.
   • CNumber of active users and session duration.
   • DCPU utilization and memory usage for each service.
   Reveal Solution Discussion Select an answer first
2. 2expert · hard

   You have a set of on-premises servers that send metrics to Cloud Monitoring via the Ops Agent. You need to be alerted if any server stops sending metrics for more than 15 minutes, but you also have servers that are intentionally decommissioned and should not trigger alerts. What should you do?

   • ACreate a dashboard that shows the last time each server sent data, and manually check it.
   • BCreate a threshold alert on the metric 'uptime' that triggers when it drops below 1.
   • CCreate a silent source detection policy for all servers, and exclude decommissioned servers by removing them from the monitored resources.
   • DCreate a log-based alert on the absence of log entries from each server.
   Reveal Solution Discussion Select an answer first
4. 3expert · hard

   You have an alerting policy that sends notifications to a Pub/Sub topic. A downstream system consumes these notifications to create incidents. You notice that some alerts are not being delivered to the downstream system, and you suspect the Pub/Sub subscription is not receiving all messages. What should you do to investigate?

   • ACheck the Cloud Monitoring dashboard to see if the alert fired.
   • BCheck the Cloud Monitoring alerting policy's notification channel configuration to ensure the Pub/Sub topic is correct.
   • CCheck the Pub/Sub subscription's metrics, such as the number of undelivered messages and the acknowledgment deadline.
   • DCreate a new alerting policy with a different notification channel to see if the issue is specific to Pub/Sub.
   Reveal Solution Discussion Select an answer first
5. 4expert · hard

   You are asked to create a dashboard that provides a comprehensive view of a critical application's health. The dashboard should be useful for both engineers and management. Engineers need to see detailed metrics, while management wants a high-level summary. What should you do?

   • ACreate two dashboards: one with detailed metrics for engineers and one with a summary for management.
   • BCreate a single dashboard with all metrics, and rely on the users to filter what they see.
   • CCreate a dashboard with only the summary metrics, and provide a link to the detailed metrics in Cloud Monitoring.
   • DCreate a dashboard with only the detailed metrics, and provide a separate summary report via email.
   Reveal Solution Discussion Select an answer first
6. 5expert · hard

   You have an alerting policy that triggers on high CPU utilization. You notice that the alert fires during a period of high traffic, but the service is still healthy. You want to reduce false alarms without missing real issues. What should you do?

   • AChange the notification channel to a less intrusive method, such as email instead of SMS.
   • BIncrease the threshold for CPU utilization to a higher value.
   • CDisable the alerting policy during periods of high traffic.
   • DAdd a condition that requires the error rate to also be high before triggering the alert.
   Reveal Solution Discussion Select an answer first