Google Professional Cloud Database Engineer Question 121
Select 3Google Cloud PlatformYou are managing a Cloud SQL instance running PostgreSQL in Google Cloud, and your team has reported degraded performance for an application connected to this database. Upon initial investigation, you suspect resource contention might be the cause. Which steps would you take to monitor and investigate the issue effectively?
- A
Check the CPU utilization metrics in the Cloud SQL monitoring dashboard.
- B
Analyze the database's RAM usage and ensure there is sufficient memory available for caching.
- C
Inspect the database's I/O operations and disk throughput in the Cloud Monitoring logs.
- D
Enable query logging to examine the slow queries and optimize them immediately.
- E
Configure a new instance with more vCPUs and memory to resolve the performance issue.
Show answer and explanation
Correct answers: A, B, C
Explanation
To investigate performance degradation in a Cloud SQL instance, monitoring key resource vitals such as CPU, RAM, and I/O is essential. These metrics help identify whether resource contention or bottlenecks are causing the issue. Query logging and scaling are secondary steps that come into play after a thorough analysis of the database vitals.
- A. Correct.
Monitoring CPU utilization is essential to identify whether the database is experiencing high CPU contention, which can lead to performance degradation.
- B. Correct.
RAM usage must be investigated to ensure there is adequate memory available for caching and overall database operations. Insufficient memory can cause heavy reliance on disk I/O, leading to slower performance.
- C. Correct.
Inspecting I/O operations and disk throughput is critical, as high I/O contention or disk bottlenecks can significantly impact database performance.
- D. Incorrect.
While enabling query logging can help identify slow queries, it is not an immediate monitoring step for resource contention. It is better suited for query optimization rather than diagnosing CPU, RAM, or I/O issues.
- E. Incorrect.
Configuring a new instance with higher resources is a reactive measure. It is not a step for monitoring or diagnosing the problem at hand.