Associate Cloud Engineer ยท Diagnostics, health and recommendations
Diagnosing application issues on Google Cloud
Cloud Trace for latency, Cloud Profiler for CPU/memory hot spots, Error Reporting, Query Insights and index advisor for databases.
Key points
- Cloud Trace shows where request latency is spent across services.
- Cloud Profiler shows which functions burn CPU or memory in production.
- Error Reporting groups exceptions so you see the top failures.
- Query Insights and the index advisor point at slow database queries and missing indexes.
Source: https://docs.cloud.google.com/trace/docs/overview
Practice questions
Which tool measures how long a request spends in each service it passes through?
- Cloud Profiler
- Cloud Trace
- Error Reporting
- Cloud NAT
Show answer
Answer: B. Cloud Trace
Cloud Trace is the distributed tracing system built for request latency. Profiler looks inside one process at CPU and memory, not across services.
Which tools help you diagnose problems inside your running application?
- Cloud Trace
- Cloud NAT
- Cloud Profiler
- Error Reporting
- Cloud Armor
Show answer
Answer: A. Cloud Trace, C. Cloud Profiler, D. Error Reporting
Trace, Profiler and Error Reporting each expose application behavior. Cloud NAT and Cloud Armor are network and security services and do not diagnose app code.
True or false: Cloud Trace keeps spans in its default bucket for 365 days.
Show answer
Answer: False
False. Spans in the _Trace bucket are kept for 30 days, so export or investigate promptly if you need older data.
Practice 8 questions on this topic in Bitzu
Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.
