1Z0-1072-25 Question 19
Single answerYou manage a mission-critical e-commerce application running on an instance pool behind a load balancer in Oracle Cloud Infrastructure. Traffic surges every day around 9 AM, and you currently use a threshold-based autoscaling policy that scales out only after average CPU utilization exceeds 85% for 10 minutes. However, by the time CPU utilization triggers the scale-out, your application has already suffered performance degradation. Which approach should you implement to ensure additional instances are provisioned before the daily spike occurs?
- A
Lower the threshold from 85% to 70% to reach the scale-out condition faster.
- B
Implement a scheduled autoscaling policy that adds instances at 8:45 AM each day.
- C
Reduce the cooldown period so that scale-outs can occur more frequently.
- D
Configure a multi-step threshold policy to immediately add two instances once CPU exceeds 85%.
Show answer and explanation
Correct answer: B
Explanation
For known, predictable traffic patterns such as daily spikes, Oracle Cloud Infrastructure supports scheduled autoscaling policies. By setting a schedule to add capacity before the anticipated surge, you avoid performance degradation caused by waiting for CPU usage to rise. For more details, refer to the official Oracle Cloud Infrastructure documentation on Autoscaling best practices.
- A. Incorrect.
Option 1 (Incorrect): Lowering the threshold to 70% might help the autoscaling policy trigger sooner, but you still rely on the load spike occurring before scaling begins. This alone may not adequately address the punctual surge at 9 AM, because the scale-out still depends on CPU metrics becoming high enough.
- B. Correct.
Option 2 (Correct): A scheduled autoscaling policy lets you proactively add instances at a specific time�in this case, just before 9 AM. This way, the infrastructure is prepared for the known surge and can handle increased traffic without performance degradation.
- C. Incorrect.
Option 3 (Incorrect): Reducing the cooldown period allows back-to-back scaling actions with less waiting time, but it does not address the root cause of being reactive. You still won�t have instances ready before the spike unless CPU usage meets the threshold first.
- D. Incorrect.
Option 4 (Incorrect): A multi-step threshold policy responds at different CPU utilization points, but it remains reactive. If usage peaks at 9 AM, you still won't have the new instances online ahead of time, risking the same performance dip.