Google Associate Cloud Engineer Question 310
Select 3Google Cloud PlatformYour team has deployed a web application on Google Cloud using Google Kubernetes Engine (GKE). Users report experiencing intermittent slow responses from the application. As a Google Certified Associate Cloud Engineer, which steps should you take to diagnose the issue using Google Cloud's diagnostic tools?
- A
Check the application logs in Google Cloud Logging for any error messages or unusual patterns.
- B
Use Google Cloud's Network Intelligence Center to analyze network latency issues.
- C
Deploy a new version of the application with increased logging verbosity.
- D
Use Google Cloud Monitoring to set up alerts for resource usage spikes such as CPU and memory.
- E
Directly connect to each GKE node to manually inspect system metrics.
Show answer and explanation
Correct answers: A, B, D
Explanation
To effectively diagnose application performance issues on Google Cloud, it's important to utilize the available cloud diagnostics tools. Google Cloud Logging and Monitoring provide insights into application logs and resource usage, while Network Intelligence Center helps analyze network issues. These tools enable you to systematically identify and address the root causes of slow responses without manual interventions.
- A. Correct.
Checking the application logs in Google Cloud Logging is essential for identifying any error messages or patterns that might indicate the source of the problem.
- B. Correct.
Google Cloud's Network Intelligence Center can be used to diagnose network-related issues, such as latency, which could be causing slow responses.
- C. Incorrect.
While increasing logging verbosity could potentially help in the future, deploying a new version without a clear understanding of the current issue is not advisable.
- D. Correct.
Setting up alerts in Google Cloud Monitoring for resource usage spikes helps identify performance bottlenecks that could contribute to slow responses.
- E. Incorrect.
Manually connecting to each GKE node is inefficient and not a best practice for diagnosing cloud-native applications.