PCDOE Optimizing performance and cost Practice Test 3 — 15 Questions

Question 1

A company runs a production workload on Google Kubernetes Engine (GKE) with cluster autoscaling enabled. The cluster has three node pools: one for general-purpose workloads (n1-standard-4), one for memory-intensive workloads (n2-highmem-8), and one for GPU-accelerated jobs (with 1 NVIDIA Tesla T4 per node). The workloads are a mix of stateless microservices and stateful databases. Over the past month, the monthly GKE cost has increased by 40% despite no significant change in application traffic or resource requests. The team has verified that vertical pod autoscaling and node auto-provisioning are enabled. They have also checked that there are no orphaned resources. They suspect that overspending is due to inefficient resource utilization or node selection. What should the team do to identify and reduce the unnecessary cost?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Use the GKE cost optimization recommender to identify idle resources and apply recommended changes.. The GKE cost optimization recommender specifically analyzes cluster utilization patterns, such as underutilized nodes or pods, and provides actionable recommendations to right-size resources. Since the team has already verified that VPA and node auto-provisioning are enabled and no orphaned resources exist, the recommender can pinpoint inefficiencies like over-provisioned node pools or idle GPU nodes that are driving the 40% cost increase despite stable traffic.

Answer

Review billing export data in BigQuery to identify the top cost contributors by project, service, and label.

Answer

Disable node auto-provisioning and switch all nodes to preemptible instances.

Answer

Create a budget alert for the GKE service and set a hard limit to prevent further overspend.

Question 2

A company runs a Cloud Run service that receives bursty traffic. During bursts, many cold starts occur and p95 latency rises. The team wants to reduce cold starts without paying for idle capacity around the clock. What should they do?

Accepted Answer

Set the minimum number of instances to a value greater than zero and configure CPU to be always allocated.. Cloud Run cold starts occur when a new instance must initialize to handle a request. Keeping a minimum instance count above zero maintains warm capacity, and always-allocated CPU avoids throttling during initialization and idle periods. This targets the burst scenario directly without the cost of running a large idle fleet continuously, because the minimum can be modest and instances are only billed while they exist.

Answer

Enable CPU throttling and reduce the container image size.

Answer

Move the service to a GKE cluster with cluster autoscaling and a horizontal pod autoscaler.

Answer

Increase the maximum number of instances and set the concurrency to 1.

Question 3

A company uses Cloud Storage to store large amounts of data that is accessed frequently for the first 30 days and then rarely accessed after that. They want to minimize storage costs while ensuring that data is immediately available when needed. What should they do?

Accepted Answer

Set a lifecycle rule to transition objects to Nearline storage after 30 days, and another rule to transition to Coldline after 90 days.. The access pattern shows frequent access for 30 days, then rare access. A tiered lifecycle policy transitions data to progressively cheaper storage classes as access frequency drops. Nearline after 30 days aligns with monthly access, and Coldline after 90 days aligns with yearly access, avoiding early deletion fees and minimizing cost.

Answer

Set a lifecycle rule to transition objects to Archive storage after 30 days.

Answer

Set a lifecycle rule to transition objects to Nearline storage after 30 days.

Answer

Set a lifecycle rule to transition objects to Coldline storage after 30 days.

Question 4

Your team runs a batch analytics job on a Dataproc cluster with 20 worker nodes. The job runs daily at 02:00 and finishes around 06:00. Job duration increased by 30% last week after the input data volume doubled, and the team wants to cut the cluster's compute spend while keeping the job within the same four-hour window. The job is shuffle-heavy and uses the default Dataproc persistent cluster. Which change should you make?

Accepted Answer

Enable Dataproc autoscaling on the cluster with a secondary worker policy so the cluster scales workers up and down based on YARN pending memory.. Autoscaling is the right lever because Dataproc can add secondary workers when YARN reports pending memory and remove them when the shuffle-heavy stages finish. Secondary workers are cheaper and can be preemptible or Spot, so spend drops during the troughs while the job still meets the four-hour window. A static commitment or a platform migration does not fix the capacity shortfall for the doubled input.

Answer

Purchase a one-year committed use discount for the existing 20-worker Dataproc cluster and leave the cluster configuration unchanged.

Answer

Convert the cluster to a Dataproc ephemeral cluster created by a workflow template, and enable graceful decommissioning with a larger worker machine type.

Answer

Move the job to a Google Kubernetes Engine cluster with cluster autoscaler enabled and run the Spark job as a Kubernetes workload.

Question 5

A company runs a stateful streaming pipeline on Dataflow that reads from Pub/Sub and writes to Bigtable. The pipeline uses a fixed number of 50 workers to handle peak load, but during off-peak hours CPU utilization is under 10%. The team wants to reduce cost while keeping the same pipeline code and without increasing data freshness. What should they do?

Accepted Answer

Enable Streaming Engine and use horizontal autoscaling with a minimum of 1 worker and a maximum of 50 workers.. Streaming Engine offloads execution from worker VMs and enables horizontal autoscaling to scale workers with backlog and CPU. A low minimum reduces off-peak cost, while a high maximum preserves peak throughput. Because the pipeline code and semantics are unchanged, data freshness is not degraded. The other options either change semantics, reduce peak capacity, or rely on a feature that does not exist for Dataflow workers.

Answer

Reduce the number of workers to 10 and increase the worker machine type to n2-standard-16.

Answer

Enable vertical autoscaling on the Dataflow workers and keep the worker count fixed at 50.

Answer

Switch the pipeline to batch mode and run it on a schedule every hour.

Question 6

A company runs a stateless web application on a managed instance group (MIG) behind an HTTP(S) Load Balancer. They want to optimize cost without sacrificing availability. The application experiences variable traffic with predictable peak hours. Which two actions should they take? (Choose two.)

Accepted Answer

Configure a committed use discount (CUD) for the baseline number of instances required during off-peak hours.. The application has variable traffic with predictable peaks. Autoscaling based on HTTP load balancing utilization ensures capacity matches demand. A CUD for the off-peak baseline reduces cost for the always-on minimum. Together, they optimize cost while maintaining availability. Other options either add complexity, risk availability, or use less effective scaling metrics.

Answer

Configure the MIG to use a custom instance template with smaller machine types during off-peak hours and larger during peak hours.

Answer

Use preemptible VMs in the MIG to reduce cost, and configure the load balancer to drain connections on preemption.

Answer

Set the autoscaler to scale based on CPU utilization with a target of 80%.

Question 7

A company uses Cloud Storage to store archival data. They want to minimize storage costs while maintaining availability. Which storage class should they use?

Accepted Answer

Archive storage class.. The Archive storage class is the correct choice because it offers the lowest storage cost for archival data that is accessed less than once per year, with a 365-day minimum storage duration and retrieval costs that are higher than other classes. This aligns with the requirement to minimize storage costs while maintaining availability, as Archive data is still available for retrieval (though with a longer latency) and is replicated for durability.

Answer

Nearline storage class.

Answer

Standard storage class.

Answer

Coldline storage class.

Question 8

A company is using BigQuery for analytics and wants to control costs. They have many queries that scan large amounts of data. Which approach is most effective in reducing query costs?

Accepted Answer

Partition tables by date and use partition pruning in queries.. Partitioning tables by date and using partition pruning in queries directly reduces the amount of data scanned by BigQuery, which is the primary driver of on-demand query costs. By filtering on the partition column, BigQuery can skip entire partitions that do not match the query criteria, minimizing the bytes processed. This is the most effective cost-control measure because it addresses the root cause of high costs—excessive data scanning—without requiring a pricing model change or additional resource commitments.

Answer

Switch to flat-rate pricing to cap costs.

Answer

Reserve BigQuery slots for dedicated capacity.

Answer

Use clustering to organize data within partitions.

Question 9

A company runs a web application on Google Kubernetes Engine (GKE) with multiple services. They want to reduce costs without impacting performance. Which THREE actions should they take? (Choose three.)

Accepted Answer

Enable node auto-provisioning to automatically adjust node pools.. Node auto-provisioning in GKE automatically creates and scales node pools based on the resource requirements of pending pods. This eliminates the need for manual node pool management and ensures that only the necessary compute resources are provisioned, reducing costs without manual intervention or over-provisioning.

Answer

Enable cluster autoscaling and manually scale nodes based on peak load.

Answer

Deploy a service mesh like Istio to optimize traffic routing.

Question 10

A DevOps team is analyzing Google Cloud costs and notices that spending on BigQuery has increased significantly. They want to reduce costs without impacting ongoing analytical workloads. Which TWO actions should they take? (Choose two.)

Accepted Answer

Set custom cost controls like query quotas and maximum bytes billed per query.. BigQuery allows you to set custom cost controls such as query quotas (e.g., concurrent queries per project) and maximum bytes billed per query. These controls cap resource usage at the query level, preventing runaway costs while still allowing analytical workloads to run within defined limits. This directly addresses cost spikes without blocking ongoing operations.

Answer

Switch to on-demand pricing to pay only for queries run.

Answer

Enable column-level security to restrict access to sensitive data.

Answer

Delete unused datasets to reduce storage costs.

Question 11

Refer to the exhibit. The output shows a recommendation from the Cloud Cost Optimization recommender for an instance in us-central1-a. The instance is a production web server that consistently runs at 25% CPU utilization during peak hours. What should the DevOps engineer do to implement this recommendation with minimal risk?

Accepted Answer

Stop the instance, change the machine type to n2-standard-4, and start it again.. The Cloud Cost Optimization recommender has identified that the current instance is over-provisioned for its actual workload (25% CPU utilization during peak hours). By stopping the instance, changing the machine type to n2-standard-4 (which has 4 vCPUs and 16 GB RAM, matching the recommended 4 vCPUs and 15 GB memory), and starting it again, the engineer rightsizes the instance without incurring downtime during the change itself. This approach minimizes risk because the instance is stopped only briefly for the machine type change, and the new machine type is a standard, well-tested configuration that aligns with the recommender's analysis.

Answer

Ignore the recommendation because the instance is production and any change might cause downtime.

Answer

Add a second n2-standard-4 instance behind a load balancer to distribute load.

Answer

Use a custom machine type with 4 vCPUs and 32 GB memory to ensure enough RAM.

Question 12

Match each SRE term to its description.

Accepted Answer

SLI: A carefully defined quantitative measure of some aspect of the level of service.. The correct matches are: SLI with a quantitative measure, SLO with a target value, SLA with a contract including consequences, and Error Budget with remaining unreliability. Common confusions include swapping Error Budget with SLI or confusing SLO with SLA.

Answer

SLI: The acceptable amount of unreliability remaining before the SLO is violated.

Answer

SLO: A contract with users that includes consequences.

Question 13

A company uses Compute Engine with committed use discounts for 1-year. They need to reduce costs further. What should they do?

Accepted Answer

Rightsize their VMs based on recommender recommendations.. Rightsizing VMs based on Google Cloud's Recommender recommendations is the correct next step because committed use discounts (CUDs) lock in a discount on a committed spend or resource amount, but if the VMs are over-provisioned, you are still paying for unused capacity. Rightsizing reduces the baseline resource consumption, which lowers both on-demand and committed costs. This is the standard FinOps practice after securing a discount.

Answer

Use sustained use discounts instead.

Answer

Use preemptible VMs for all workloads.

Answer

Increase committed use discount term to 3 years.

Question 14

A team wants to receive a notification when their monthly spending exceeds 80% of the budget. Which GCP feature should they configure?

Accepted Answer

Budget alert with threshold rule at 0.8.. GCP Budgets allow you to set a budget amount and configure alert thresholds expressed as a fraction of the budget (e.g., 0.8 for 80%). When actual or forecasted spend crosses the threshold, Budget Alert notifications are sent via email or Pub/Sub, enabling the team to react before the budget is exceeded.

Answer

Committed use discounts.

Answer

Cloud Storage bucket for cost exports.

Answer

Cloud Monitoring cost insights dashboard.

Question 15

A DevOps engineer notices that the monthly compute cost is higher than expected for a Kubernetes Engine cluster. They want to identify which namespaces or workloads are contributing most to the cost. What should they do?

Accepted Answer

Enable GKE Usage Metering and export to BigQuery.. GKE Usage Metering exports detailed per-cluster resource consumption data (CPU, memory, storage, and network) to BigQuery, enabling cost attribution by namespace, label, or workload. This is the only option that directly provides granular cost breakdowns for Kubernetes resources, allowing the engineer to identify which namespaces or workloads drive compute costs.

Answer

Enable Logging for all containers and analyze logs.

Answer

Use Cloud Monitoring dashboards with Kubernetes metrics.

Answer

Add labels to nodes and use Cost Table reports.