Google Professional Cloud DevOps Engineer
Get started today
Ultimate access to all questions.
Ultimate access to all questions.
Your company processes IoT data at scale using Pub/Sub, App Engine standard environment, and a Go application. During peak load, performance degrades inconsistently, but you cannot reproduce this issue locally. You need to continuously monitor the production application to identify slow code paths while minimizing performance impact and management overhead. What should you do?
Explanation:
To continuously monitor the application in production with minimal performance impact and management overhead, the best approach is to use Cloud Profiler. Cloud Profiler is a fully managed service that provides continuous profiling for applications running on Google Cloud, including those on App Engine. It helps identify slow paths in the code by collecting CPU and heap profiles with low overhead. Initializing the cloud.google.com/go/profiler library in the application enables this integration. Option A, using Cloud Monitoring to assess CPU utilization, does not provide code-level insights into performance issues. Option B, installing a continuous profiling tool into Compute Engine, introduces unnecessary management overhead and is not as seamless as using Cloud Profiler. Option C, periodically running the go tool pprof command, does not provide continuous monitoring and requires manual intervention. Therefore, the correct answer is D.