Your company runs one batch process in an on-premises server that takes around 30 hours to complete. The task runs monthly, can be performed offline, and must be restarted if interrupted. You want to migrate this workload to the cloud while minimizing cost. What should you do?
For a batch process that takes 30 hours to complete, is not fault-tolerant, and must be restarted if interrupted, migrating the workload to a Compute Engine VM (option B) is the most reliable and cost-effective solution. This allows you to start and stop the instance as needed, ensuring the process is completed without interruption, as Spot VMs can be preempted at any time. This method provides flexibility, control over the execution time, and avoids the risk associated with potential interruption of Spot VMs.
B. Migrating the workload to a Compute Engine VM and starting and stopping the instance as needed allows you to control when the task runs. This approach provides flexibility in terms of when to initiate the batch process, and it can be easily scheduled to run monthly. By stopping the instance when the task is not running, you can save on compute costs.
I had this question in my exam. There was Spot VM option available, but I still picked manual starting/stopping as job is not fault tolerant.
D https://cloud.google.com/compute/docs/instances/spot#limitations Spot VMs are unlike Preemptible VMs (https://cloud.google.com/compute/docs/instances/preemptible#limitations) , which do not stop after 24 hours. And we can re-start the batch job in case if it gets stopped in between as per the question. So D is the best cost optimal solution.
Answer B. Spot VMs are not suitable for batch processes that take an extended period of time to complete, such as a 30-hour batch process. Spot VMs typically have a maximum runtime of 24 hours.
Spot VM runs until it is stopped (by google), Unlike preemptible, which has a lifetime 24 hr
another repeated question.
I don't believe Spot VMs are the best solution here since they can be preempted by Google at any time. Option B prioritizes predictable execution and cost optimization for this scheduled batch process. It balances cost-savings with the need for guaranteed completion, aligning well with the provided scenario.
wonderfully said.
An Instance Template with Spot VMs in a Managed Instance Group (MIG) has no 24-hour limitation, can restart via the MIG if interrupted, and offers low costs, making it a perfect match.
ok, but the questions says that the process can NOT be interruputed, if does, you must restarted from the begining...
Must be B. Stop instances are cost-effective yes and are for batch jobs. HOWEVER, after 24h teh instances will be change and, thus, the job will restart again and again after 24h. The only option that dont use Stop instances are the B. So, my answer is B.
Only preemptive VMs have that 24h time limit, the newer version Spot VMs don't have the time limit. But answer is still B because, they are not fault tolerant
It seems to me that answer is A. Scale to zero - After completed batch processes, minimize cost. Spot VM - minimize cost. failure resistant - do not need restart instance. You already have a template - delete and deploy your Compute instances once a month easily and fast. Can you explain me unless I understand some thing.
A is the most robust and cost-effective for your scenario. This setup leverages the cost savings of Spot VMs while also providing the ability to handle interruptions through a Managed Instance Group, ensuring the process is completed even if individual instances are preempted. Additionally, the automatic scaling based on CPU utilization helps manage resources efficiently, further reducing costs. Thus, A not only minimizes costs by using Spot VMs but also enhances reliability and scalability through a Managed Instance Group, making it the most suitable choice for migrating your batch process to the cloud.
"runs one batch process" So, why would you like a instange group? it is only one batch process. And if it is interrupted, it has to be restarted...for me it is B.
It seems to me that answer is A. Scale to zero - After completed batch processes, minimize cost. Spot VM - minimize cost. failure resistant - do not need restart instance. You already have a template - delete and deploy your Compute instances once a month easily and fast. Can you explain me unless I understand some thing.
I don´t like this kind of questions... but I think is A.. An Instance Template with Spot VMs (not preemptible)...
Same as 136
no, this question has different answer choices with the inclusion of spot vms
Spot VM best practices
Spot VM best practices
Spot VM best practices
Per ChatGPT, Option A is the most appropriate solution as it leverages the cost benefits of Spot VMs, provides scalability through Managed Instance Groups, and aligns with the characteristics of the workload, resulting in minimized cost for running the batch process in the cloud.
https://cloud.google.com/compute/docs/instances/create-use-spot#best-practices Spot VM best practices create an instance template and Spot VM does not have a minimum or maximum runtime to there is no 24 hour limit as suggested by some.
Read the link you shared, it says: "Spot VMs are virtual machine (VM) instances with the spot provisioning model. Spot VMs are available at a 60-91% discount compared to the price of standard VMs. However, Compute Engine might reclaim the resources by preempting Spot VMs at any time. Spot VMs are recommended only for fault-tolerant applications that can withstand VM preemption." The last part literally says that Spot VMs are recommended for fault-tolerant appliactions, by the questions, this application is not fault tolerant at all mate.