MENU
A team of data scientists infrequently needs to use a Google Kubernetes Engine (GKE) cluster that you manage. They require GPUs for some long-running, non-restartable jobs. You want to minimize cost. What should you do?
Suggested Answer: B
Get Premium Associate Cloud Engineer Questions
as Interactive Web-Based Practice Test or PDF
Real Exam Format
Detailed Explanations
Instant Access
Get Your Practice Test
37 comments · newest first
LE
Lenna
11 months ago
VerticalPodAutoscaler is cool, but not for GPU jobs, right?
LE
Lenny
11 months ago
Node auto-provisioning might not save enough costs.
FR
France
11 months ago
Surprised that preemptible VMs can handle long-running jobs.
MO
Mona
11 months ago
I think autoscaling is a better option for flexibility.
LI
Ling
12 months ago
Preemptible VMs are way cheaper!
ME
Melissia
12 months ago
I'm a bit confused about the VerticalPodAutoscaler; I thought it was more for optimizing resource requests rather than cost-saving.
EA
Earleen
1 year ago
I practiced a similar question where we had to choose between autoscaling and preemptible VMs. I feel like option C might be the right choice.
AR
Arleen
1 year ago
I think enabling node auto-provisioning could help, but it might not be the most cost-effective for infrequent jobs.
CA
Carol
1 year ago
I remember we discussed using preemptible VMs to save costs, but I'm not sure if that's the best option here.
MA
Matthew
1 year ago
I'm pretty confident that option C is the best choice here. Preemptible VMs with GPUs should give us the cost savings we're looking for while still meeting the data scientists' requirements.
DO
Donte
1 year ago
Hmm, I'm a bit unsure about this one. I'm leaning towards option D, as it seems to provide the flexibility of autoscaling the node pool with GPUs. But I'll need to think it through a bit more.
RE
Reita
1 year ago
This seems like a tricky question, but I think option C is the way to go. Using preemptible VMs with GPUs should help minimize costs while still providing the necessary resources.
DO
Domingo
1 year ago
Option A with node auto-provisioning sounds like a good way to automatically scale up the cluster as needed. That could be a simple and effective solution to the problem.
SA
Sarah
1 year ago
Hmm, I'm a little unsure about this one. I know the question is about workforce education, but I'm not totally confident that Security Awareness and Training is the right answer. I might need to re-read the question and options a few times to make sure I'm picking the best choice.
JU
Justine
1 year ago
This seems like a tricky one, but I'll give it my best shot. I'll start by eliminating any options that don't seem to fit the description of corrosion from oxygen.
PH
Phyliss
1 year ago
Okay, I remember learning about VLANs in class. I believe the key is that they create logical network segments, which is what option B describes. I'm pretty confident that's the right answer.
HO
Howard
1 year ago
I remember it has something to do with maintaining secure connections, but I keep mixing up how it relates to WAN Edge routers and vSmart.
DO
Dorothea
1 year ago
If I had a penny for every time I've had to optimize for cost on a certification exam, I'd be rich enough to buy my own GKE cluster.
LY
Lynna
1 year ago
A) Enable node auto-provisioning on the GKE cluster.
DO
Dorian
1 year ago
D) Create a node pool of instances with GPUs, and enable autoscaling on this node pool with a minimum size of 1.
KA
Karima
1 year ago
C) Create a node pool with preemptible VMs and GPUs attached to those VMs.
JU
Justine
1 year ago
Wait, there's an option to use preemptible VMs with GPUs? That's like the holy grail of cost optimization!
NG
Ngoc
1 year ago
I didn't know that was an option, thanks for pointing it out!
CA
Carman
1 year ago
That's right! It's a great way to minimize costs while still getting the GPU power you need.
GE
Georgene
1 year ago
Yes, option C allows you to create a node pool with preemptible VMs and GPUs attached to those VMs.
RO
Rozella
1 year ago
A and B don't really address the GPU requirement, so I'm leaning towards C. Gotta love those preemptible VMs!
NI
Nikita
2 years ago
D sounds tempting, but I'd rather not have to worry about scaling up and down. C is the clear winner here.
CH
Chan
1 year ago
It might be more convenient, but C is still the most cost-effective choice.
KR
Kris
1 year ago
But wouldn't D be more convenient in terms of autoscaling?
AR
Arlene
1 year ago
I agree, using preemptible VMs with GPUs attached seems like the most cost-effective solution.
LO
Lorriane
1 year ago
I think C is the best option for minimizing cost.
NA
Narcisa
2 years ago
That's a good point. Autoscaling with GPUs could be more cost-effective in the long run.
AL
Alisha
2 years ago
I disagree, I believe option D is better. Autoscaling with GPUs can ensure the jobs are completed efficiently.
AL
Allene
2 years ago
Hmm, I think C is the way to go. Preemptible VMs with GPUs? That's a cost-saver for sure!
BA
Barabara
1 year ago
Yes, it's a great way to minimize cost while still meeting the data scientists' needs.
TA
Tamar
1 year ago
I agree, using preemptible VMs with GPUs attached is a cost-effective solution.
NA
Narcisa
2 years ago
I think option C is the best choice. Preemptible VMs with GPUs can help minimize cost.