Question 1
IntermediatePerformance Management · Identify the metrics that should be used to investigate a performance issue
A performance analyst notices that several tactical queries are experiencing high and variable response times. Using the ResUsage tables, the analyst wants to determine if the issue is caused by AMP-level CPU contention rather than I/O waits or delays. Which combination of ResUsage metrics provides the most direct evidence of CPU queuing at the AMP level?
Answer and explanation
Correct answer: C
CPUServ represents the CPU seconds used by tasks for the AMP. A high value is expected on a busy system. However, the key indicator for CPU contention is the RunQ, which represents the average number of tasks waiting for CPU resources on an AMP. A consistently high RunQ value alongside high CPUServ directly indicates that more work is arriving than the CPU can process, leading to queuing and increased response times. The other options mix CPU metrics with I/O or delay metrics, which are less direct for diagnosing CPU-specific queuing.
