Exam Professional Cloud Architect topic 1 question 199 discussion

Actual exam question from Google's Professional Cloud Architect

Question #: 199
Topic #: 1

[All Professional Cloud Architect Questions]

You have a Compute Engine application that you want to autoscale when total memory usage exceeds 80%. You have installed the Cloud Monitoring agent and configured the autoscaling policy as follows:

You observe that the application does not scale under high load. You want to resolve this. What should you do?

A. Change the Target type to DELTA_PER_MINUTE.
B. Change the Metric identifier to agent.googleapis.com/memory/bytes_used.
C. Change the filter to metric.label.state = ‘used’.
D. Change the filter to metric.label.state = ‘free’ and the Target utilization to 20.

Show Suggested Answer

Suggested Answer: C 🗳️

by JamesKarianis at Aug. 27, 2024, 7:09 p.m.

Comments

Submit Cancel

Nick89GR

Highly Voted 9 months, 1 week ago

I am confused since this question comes in contrast with the answer in question 190. Anyone knows what is the real answer? I would expected used+buffered+cached+slab Which means the sum of all these.

upvoted 13 times

Piddi

4 months, 1 week ago

yes..it is confusing. I used Claude.ai to clarify my doubt. - By monitoring metric.label.state = 'free' memory, you're tracking available memory instead of used memory - Setting the target to 20% means autoscaling will trigger when free memory drops below 20% (equivalent to total usage exceeding 80%) - This approach accurately captures all memory usage regardless of how it's being utilized by the system

upvoted 1 times

...

JamesKarianis

Highly Voted 11 months, 1 week ago

C. Change the filter to metric.label.state = ‘used’. The current filter is set up with multiple AND conditions, which means it's looking for a metric that simultaneously has all these states: 'used', 'buffered', 'cached', and 'slab'. This is logically impossible, as a memory location can't be in multiple states at once. Therefore, the filter will never match any metrics, and the autoscaling policy won't trigger.

upvoted 10 times

...

frank_tsai_tech

Most Recent 4 months, 2 weeks ago

Selected Answer: C

Explanation: • Issue with the Current Filter: The filter you configured uses multiple conditions with the AND operator for states (‘used’, ‘buffered’, ‘cached’, ‘slab’). However, each time series reported by the Cloud Monitoring agent includes a single state label value. This means no single data point can satisfy all these conditions simultaneously. • Correcting the Filter: To monitor total memory usage accurately, you should only focus on the “used” memory state. Changing the filter to just metric.label.state = 'used' ensures that the autoscaler is correctly evaluating the actual used memory percentage. Once the used memory exceeds the target (80%), the autoscaler will trigger scaling. Thus, modifying the filter as in option C resolves the issue.

upvoted 2 times

...

JaquiMB

7 months ago

Selected Answer: C

Autoscale based on memory usage To configure autoscaling based on the percent of used memory, specify the percent_used metric provided by the memory Ops Agent metrics. You should filter the metric by state to use only the used memory state. If you do not specify the filter, then the autoscaler takes the sum of memory usage by all memory states labeled as buffered, cached, free, slab, and used.

upvoted 3 times

...

tangac

11 months, 1 week ago

Selected Answer: C

as it is clearly indicated in the public documentation https://cloud.google.com/compute/docs/autoscaler/scaling-cloud-monitoring-metrics#gcloud_5 you have change the filter to metric.label.state="used"

upvoted 9 times

...