Google Professional Cloud DevOps Engineer Question 138
Select 3Google Cloud PlatformYour team has deployed a new application on Google Kubernetes Engine (GKE) and you have started receiving alerts about high memory usage on some of the pods. You want to investigate the issue further by performing ad hoc analysis of memory usage across the cluster. Using Metrics Explorer in Google Cloud Monitoring, which steps should you take to identify the pods with the highest memory usage?
- A
Select the 'Kubernetes Container' resource type and filter by 'Memory Usage' metric.
- B
Use the 'Group By' feature in Metrics Explorer to group data by namespace and pod name.
- C
Apply a filter in Metrics Explorer to only display metrics for pods with memory usage above a certain threshold.
- D
Switch to the 'Custom Dashboards' section in Cloud Monitoring to analyze memory usage.
- E
Export the metrics data to BigQuery for advanced analysis.
Show answer and explanation
Correct answers: A, B, C
Explanation
Metrics Explorer is a powerful tool in Google Cloud Monitoring for performing ad hoc analysis of metrics. In this scenario, selecting the correct resource type and metric, grouping data by relevant dimensions, and applying filters help identify pods with high memory usage effectively. While other tools like custom dashboards and BigQuery have their use cases, they are not appropriate for quick and interactive troubleshooting in this context.
- A. Correct.
Correct: 'Kubernetes Container' is the appropriate resource type for analyzing metrics at the container level in GKE, and 'Memory Usage' is the relevant metric for this scenario.
- B. Correct.
Correct: Grouping by namespace and pod name allows you to isolate memory usage patterns for specific pods, which is essential for identifying those with the highest memory usage.
- C. Correct.
Correct: Applying a filter to show only pods exceeding a specific memory threshold helps narrow down the results to problematic pods.
- D. Incorrect.
Incorrect: While custom dashboards are useful for monitoring predefined metrics, they are not ideal for ad hoc analysis, which is better suited for Metrics Explorer.
- E. Incorrect.
Incorrect: Exporting to BigQuery is suitable for long-term or complex analysis but is unnecessary for quick ad hoc analysis using Metrics Explorer.