MENU
You company has a Kubernetes application that pulls messages from Pub/Sub and stores them in Firestore. Because the application is simple, it was deployed as a single pod. The infrastructure team has analyzed Pub/Sub metrics and discovered that the application cannot process the messages in real time. Most of them wait for minutes before being processed. You need to scale the elaboration process that is I/O-intensive. What should you do?
Suggested Answer: A
Get Premium Professional Cloud Architect (PR000213) Questions
as Interactive Web-Based Practice Test or PDF
Real Exam Format
Detailed Explanations
Instant Access
Get Your Practice Test
35 comments · newest first
JU
Julene
11 months ago
Surprised that they didn't mention horizontal pod autoscaling earlier!
RE
Rebbecca
11 months ago
I’m not sure about D, it seems too specific for this scenario.
FA
Fatima
11 months ago
Wait, isn't option A also a good option?
IN
Ines
11 months ago
Totally agree, C makes the most sense here!
JA
Javier
12 months ago
I think option C is the best choice for scaling based on undelivered messages.
ZO
Zoila
12 months ago
I believe option D sounds familiar, but I’m not entirely confident if setting CPU percentage is the right approach for an I/O-bound application.
FI
Filiberto
1 year ago
I vaguely recall that using the --enable-autoscaling flag is more about enabling the feature rather than configuring it based on specific metrics.
EM
Emelda
1 year ago
I think we practiced a similar question where we had to choose between different metrics for scaling. I feel like subscription/num_undelivered messages might be relevant here.
VA
Valentine
1 year ago
I remember we discussed using metrics for autoscaling, but I'm not sure which one is best for I/O-intensive tasks.
LA
Latrice
1 year ago
Wait, what's the difference between the subscription/push_request and subscription/num_undelivered message metrics? I'm not sure which one would be more appropriate for this scenario. Maybe I should just go with the kubectl autoscale command in option D to keep it simple.
VA
Vanesa
1 year ago
Okay, I've got this. The key is to use the Pub/Sub subscription/push_request metric to trigger the autoscaling. That way, the Kubernetes cluster will automatically scale up and down based on the incoming message load. Option A is the way to go.
KR
Kris
1 year ago
This seems like a straightforward Kubernetes autoscaling question. I think I'll go with option C and configure the autoscaling based on the Pub/Sub num_undelivered message metric.
WY
Wynell
1 year ago
Hmm, I'm a bit confused here. Do I need to enable autoscaling at the cluster level or can I just configure it for the specific deployment? I'll have to review the Kubernetes autoscaling documentation again.
NO
Norah
1 year ago
Hmm, I'm a bit unsure about some of the options here. I'll need to review my notes on SOAP web services to make sure I'm selecting the right standards.
AN
Anabel
2 years ago
I'm picturing the poor Kubernetes cluster trying to process all those messages like a hamster on a wheel. Option C for the win!
LO
Louvenia
2 years ago
Wait, there's a --enable-autoscaling flag? That sounds too good to be true. I'm sticking with C, it's the most straightforward way to scale based on the actual problem.
LA
Lashanda
2 years ago
I see your point, but I still think C is the most straightforward option to address the issue.
OR
Orville
2 years ago
I disagree, I believe D is the way to go. It allows you to configure the autoscaling parameters easily.
ST
Stephen
2 years ago
I think A is the best option, it scales based on the actual message processing metric.
TA
Tatum
2 years ago
Option A is tempting, but I think C is the better choice. I don't want to end up with a bunch of pods that can't keep up with the message load.
ER
Erinn
2 years ago
A) Configure a Kubernetes autoscaling based on the subscription/num_undelivered message metric.
TY
Tyra
2 years ago
C) Configure a Kubernetes autoscaling based on the subscription/num_undelivered message metric.
NO
Noah
2 years ago
A) Configure a Kubernetes autoscaling based on the subscription/num_undelivered message metric.
LI
Lizette
2 years ago
That's a good point, but I still think configuring based on the num_undelivered message metric would be more effective.
WI
Wilford
2 years ago
I disagree, I believe we should use kubectl autoscale deployment APP_NAME --max 6 --min 2 --cpu-percent 50 to configure autoscaling.
LA
Latosha
2 years ago
Haha, I bet the infrastructure team is pulling their hair out trying to keep up with all those messages! Option D looks good, but I'd probably go with C to be safe.
IL
Ilona
2 years ago
Yeah, it's a tough call between Option C and Option D. Both could potentially solve the issue.
TI
Tiffiny
2 years ago
I think Option D could also work well, but I see your point about going with Option C for safety.
LA
Lacey
2 years ago
I agree, the infrastructure team must be stressed out. Option C does seem like a safer choice.
NA
Naomi
2 years ago
Option C seems like the way to go. Scaling based on the num_undelivered message metric makes sense for an I/O-intensive application.
PA
Pansy
2 years ago
I agree, scaling based on the num_undelivered message metric makes perfect sense.
SH
Shaniqua
2 years ago
Option C is definitely the best choice for an I/O-intensive application.
ME
Melvin
2 years ago
C) Configure a Kubernetes autoscaling based on the subscription/num_undelivered message metric.
MI
Mica
2 years ago
A) Configure a Kubernetes autoscaling based on the subscription/num_undelivered message metric.
LI
Lizette
2 years ago
I think we should configure Kubernetes autoscaling based on the subscription/num_undelivered message metric.