Google ACE Deploying and Implementing a Cloud Solution Practice Test 3 — 15 Questions

Question 1

A company wants to deploy a stateless microservice on GCP. The service experiences unpredictable traffic bursts and must scale to zero when idle to minimize costs. Which two services meet these requirements?

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

Accepted Answer

Cloud Functions (HTTP trigger). Cloud Run (fully managed) and Cloud Functions (HTTP trigger) both scale to zero when idle and handle burst traffic. Additionally, App Engine Standard environment can also scale to zero when configured with manual or basic scaling, or with automatic scaling and min_idle_instances set to 0. Compute Engine managed instance groups can scale down but not to zero as they require at least one instance. GKE with managed instance groups also requires a minimum number of nodes. Therefore, the correct answers are Cloud Functions, Cloud Run, and App Engine Standard.

Answer

Compute Engine managed instance group with autoscaling

Answer

Google Kubernetes Engine (GKE) with a managed instance group

Question 2

An engineer needs to create a Compute Engine instance with a specific custom subnet, an Ubuntu 20.04 LTS image, and a 50 GB boot disk. The engineer also wants to run a startup script that installs Apache. Which gcloud command should the engineer use to create this instance?

Accepted Answer

gcloud compute instances create my-instance --machine-type=e2-medium --image-family=ubuntu-2004-lts --image-project=ubuntu-os-cloud --boot-disk-size=50GB --subnet=my-subnet --zone=us-central1-a --metadata startup-script='apt-get update && apt-get install -y apache2'. It uses the proper gcloud syntax for creating a Compute Engine instance with a custom subnet, Ubuntu 20.04 LTS image, 50 GB boot disk, and a startup script. The --machine-type flag is correctly named (not --machine), the --image-family flag is used with the correct family ubuntu-2004-lts, and the --metadata flag uses a space-separated key-value pair (startup-script='...') which is the required format for metadata. The --boot-disk-size=50GB and --subnet=my-subnet are also valid and correctly specified.

Answer

gcloud compute instances create my-instance --machine=e2-medium --image-family=ubuntu-2004-lts --image-project=ubuntu-os-cloud --boot-disk-size=50GB --subnet=my-subnet --zone=us-central1-a --metadata=startup-script='apt-get update && apt-get install -y apache2'

Answer

gcloud compute instances create my-instance --machine-type=e2-medium --image-family=ubuntu-1804-lts --image-project=ubuntu-os-cloud --boot-disk-size=50GB --subnet=my-subnet --zone=us-central1-a --metadata=startup-script='apt-get update && apt-get install -y apache2'

Answer

gcloud compute instances create my-instance --machine-type=e2-medium --image=ubuntu-2004-lts --image-project=ubuntu-os-cloud --boot-disk-size=50GB --subnet=my-subnet --zone=us-central1-a --metadata startup-script='apt-get update && apt-get install -y apache2'

Question 3

An engineer is deploying a containerized application on Google Kubernetes Engine (GKE) and needs to expose it to the internet with a static external IP address. The cluster is in a VPC-native configuration. Which approach should the engineer take to meet this requirement?

Accepted Answer

Reserve a static external IP address in the same region as the cluster, then create a Service of type LoadBalancer and assign the static IP via the loadBalancerIP field.. To expose a service with a static external IP on GKE, the engineer should reserve a static external IP address, then create a Service of type LoadBalancer and specify that IP address in the service spec using the loadBalancerIP field. Option C correctly describes this. Option A is incorrect because a Service of type NodePort does not provide an external IP on its own; combining it with Cloud NAT does not expose the service with a static external IP. Option B is incorrect because ClusterIP is internal only and cannot be combined with Cloud Load Balancing for a static external IP. Option D is incorrect because while an Ingress can expose a service, it typically requires additional configuration for a static IP and the described approach with NodePort is not straightforward.

Answer

Create a Service of type NodePort and then use a Cloud NAT with a static IP.

Answer

Create a Service of type ClusterIP and then use a Cloud Load Balancing service with the static IP.

Answer

Create a Service of type NodePort and then create an Ingress resource with a static IP annotation.

Question 4

A company wants to store backup files that are accessed only a few times per year, with retrieval times of up to a few hours. The files must be stored at the lowest possible cost. Which Cloud Storage storage class should the company use?

Accepted Answer

Archive. The Archive storage class is the lowest-cost option for long-term storage of data accessed less than once a year, with retrieval times in hours. It is ideal for backups and archival data. Standard is for frequently accessed data. Nearline is for data accessed once a month. Coldline is for data accessed once a quarter.

Answer

Coldline

Answer

Nearline

Answer

Standard

Question 5

A team is using Terraform to manage Google Cloud resources. They want to store the Terraform state file in a Cloud Storage bucket to enable collaboration. The bucket is in project 'state-project'. Which backend configuration should they use?

Accepted Answer

terraform { backend "gcs" { bucket = "state-project-tfstate" prefix = "prod" } }. The correct backend configuration for storing Terraform state in a Google Cloud Storage bucket is backend "gcs" with the bucket name and optional prefix. The bucket name must be globally unique, and the prefix allows organizing state files within the bucket. This configuration enables remote state storage and collaboration.

Answer

terraform { backend "cloud" { bucket = "my-bucket" } }

Answer

terraform { backend "remote" { hostname = "app.terraform.io" organization = "my-org" workspaces { name = "prod" } } }

Answer

terraform { backend "s3" { bucket = "my-bucket" key = "prod/terraform.tfstate" region = "us-east1" } }

Question 6

An administrator needs to create a Cloud Storage bucket in the US with the Nearline storage class and make objects publicly readable via IAM. Which gcloud command should they use to create the bucket?

Accepted Answer

gsutil mb -c NEARLINE -l US gs://my-bucket. The correct command is `gsutil mb -c NEARLINE -l US gs://my-bucket`. This creates a bucket with the Nearline storage class in the US multi-region. Option B uses an invalid flag `-s` (the correct flag for storage class is `-c`) and specifies a regional location (US-CENTRAL1) instead of a multi-region. Option C uses `gcloud storage buckets create` with `--storage-class`, which is not a valid flag; the correct gcloud flag is `--default-storage-class`. Option D uses `gcloud storage buckets create` with `--default-storage-class` but specifies the location as `us` (lowercase), which is invalid; the correct location for a multi-region bucket in the US is `US` (uppercase). Therefore, only command A correctly creates a bucket meeting the requirements.

Answer

gsutil mb -s NEARLINE -l US-CENTRAL1 gs://my-bucket

Answer

gcloud storage buckets create gs://my-bucket --storage-class=NEARLINE --region=US

Answer

gcloud storage buckets create gs://my-bucket --default-storage-class=NEARLINE --location=us

Question 7

An engineer deployed a container to Cloud Run but the service is not receiving traffic. The engineer checks the logs and finds no errors. The container listens on port 8080. What is the most likely cause of the issue?

Accepted Answer

The container is listening on port 3000 instead of the port Cloud Run expects.. Cloud Run injects a PORT environment variable (default 8080) and expects the container to listen on that port. If the container listens on a different port, such as 3000, Cloud Run will not route traffic to it. Option A is incorrect because Cloud Run supports deploying from source using buildpacks. Option C is incorrect because authentication settings do not affect traffic routing; they control access. Option D is incorrect because Cloud Run can handle concurrent requests, but a container that doesn't support concurrency will still receive traffic, just one at a time. Therefore, the most likely cause is a port mismatch.

Answer

The container was deployed from source and Cloud Run requires a pre-built image.

Answer

The service requires authentication but --allow-unauthenticated was not set.

Answer

The container does not support concurrent requests.

Question 8

A company is migrating a legacy monolithic application to GKE. The application consists of multiple microservices that need to communicate with each other. The team wants to manage traffic routing, implement canary deployments, and provide SSL termination. Which three Google Cloud services should they consider using together? (Choose three.)

Accepted Answer

Cloud Endpoints. Cloud Load Balancing provides SSL termination and distributes traffic to backend services, enabling traffic routing and canary deployments by adjusting weights between versions. Cloud CDN caches content, reducing latency and offloading traffic, beneficial for microservices. Cloud Endpoints provides API management, including traffic routing, versioning, and canary releases for APIs. Together, these three services address the requirements. Cloud NAT (for outbound connectivity) and Cloud Armor (security) do not provide the required functionality.

Answer

Cloud NAT

Answer

Cloud Armor

Question 9

An engineer is deploying a Cloud Function that processes files uploaded to a Cloud Storage bucket. The function needs to be triggered by new object creation events. The engineer has already written the function code. Which commands should the engineer run to create the bucket and deploy the function with the correct trigger? (Choose three.)

Accepted Answer

gcloud functions deploy my-function --runtime python39 --trigger-bucket my-bucket --entry-point main --region=us-central1. To create a Cloud Storage bucket, you can use either the `gsutil mb` command (option D) or the `gcloud storage buckets create` command (option E); both are valid. To deploy a Cloud Function triggered by Cloud Storage events, use `gcloud functions deploy` with the `--trigger-bucket` flag (option A). Option B is incorrect because it uses `--trigger-http` (HTTP trigger), which does not match the requirement. Option C is incorrect because it omits the trigger flag entirely. The task requires two distinct commands: one to create the bucket and one to deploy the function with the correct trigger. Therefore, the question should ask to choose two commands, for example, A and D.

Answer

gcloud functions deploy my-function --runtime python39 --trigger-http --entry-point main --region=us-central1

Answer

gcloud functions deploy my-function --runtime python39 --entry-point main --region=us-central1

Answer

gcloud storage buckets create gs://my-bucket --location=us-central1

Question 10

An engineer needs to create a Compute Engine instance with 4 vCPUs, 16 GB of memory, and a 100 GB boot disk using the Ubuntu 22.04 LTS image. Which gcloud command accomplishes this?

Accepted Answer

gcloud compute instances create my-vm --zone=us-central1-a --machine-type=n1-standard-4 --image-family=ubuntu-2204-lts --image-project=ubuntu-os-cloud --boot-disk-size=100GB. The correct command is Option C. It uses machine-type n1-standard-4 (4 vCPU, 15 GB memory, closest to 16 GB), image-family ubuntu-2204-lts (Ubuntu 22.04 LTS), image-project ubuntu-os-cloud (official Ubuntu image project), and boot-disk-size 100GB. Option A uses ubuntu-2004-lts (Ubuntu 20.04 LTS, not 22.04). Option B uses n1-highmem-4 (4 vCPU, 26 GB memory) which exceeds the memory requirement, and uses ubuntu-minimal-2204-lts (minimal image, not standard). Option D uses e2-standard-4 (4 vCPU, 16 GB memory) but image-project debian-cloud is incorrect (should be ubuntu-os-cloud).

Answer

gcloud compute instances create my-vm --zone=us-central1-a --machine-type=n1-standard-4 --image-family=ubuntu-2004-lts --boot-disk-size=100GB

Answer

gcloud compute instances create my-vm --zone=us-central1-a --machine-type=n1-highmem-4 --image-family=ubuntu-minimal-2204-lts --image-project=ubuntu-os-cloud --boot-disk-size=100GB

Answer

gcloud compute instances create my-vm --zone=us-central1-a --machine-type=e2-standard-4 --image-family=ubuntu-2204-lts --image-project=debian-cloud --boot-disk-size=100GB

Question 11

A company has a managed instance group (MIG) of web servers behind a load balancer. They want to deploy a new version of the application with zero downtime and the ability to roll back quickly if issues arise. Which update strategy should they use?

Accepted Answer

Create a new instance template with the updated application, then perform a rolling update on the MIG using the canary update strategy with a target of 10% for the new template. Validate and then promote to 100%.. A canary rolling update on a MIG lets you deploy the new instance template to a small percentage (e.g., 10%) of instances first, validate application health, and then promote the update to 100%—all while the load balancer continues serving traffic from the remaining instances. If problems arise, you can roll back to the previous template quickly. This satisfies both zero-downtime and fast rollback requirements.

Answer

Delete the MIG and recreate it with the new instance template.

Answer

Perform a rolling update on the MIG with maxSurge=0 and maxUnavailable=1 to replace instances one by one.

Answer

Create a new MIG with the new template and a separate load balancer, then update DNS to point to the new load balancer.

Question 12

An organization needs to deploy a Cloud Function that processes messages from a Pub/Sub topic and writes results to BigQuery. The function must be triggered by every message published to the topic. Which deployment command is correct?

Accepted Answer

gcloud functions deploy process-messages --runtime python310 --trigger-topic my-topic --entry-point process --region us-central1 --memory 256MB. The correct command uses --trigger-topic my-topic, runtime python310, entry point process, region us-central1, and memory 256MB. Option B is correct because it includes all required parameters. Option A is incorrect because it uses --trigger-http instead of --trigger-topic. Option C is incorrect because it omits the --memory parameter, which is required for Cloud Functions (2nd gen) or recommended for consistent performance. Option D is incorrect because it uses runtime nodejs18, but the function is written in Python as indicated by the entry point and typical exam context.

Answer

gcloud functions deploy process-messages --runtime python310 --trigger-http --entry-point process --region us-central1 --memory 256MB

Answer

gcloud functions deploy process-messages --runtime python310 --trigger-topic my-topic --entry-point process --region us-central1

Answer

gcloud functions deploy process-messages --runtime nodejs18 --trigger-topic my-topic --entry-point process --region us-central1 --memory 256MB

Question 13

A developer needs to upload a large number of files from their local machine to a Cloud Storage bucket. Which command should they use?

Accepted Answer

gsutil cp /local/dir/* gs://my-bucket/. The correct command to upload files to a Cloud Storage bucket is `gsutil cp`. Option B is correct. Option A (`gcloud storage ls`) lists objects in a bucket, not uploads. Option C (`gcloud storage buckets create`) creates a new bucket. Option D (`gsutil rsync -r`) synchronizes directories, which is more complex and not necessary for a simple upload of a large number of files; `gsutil cp` is the standard upload command.

Answer

gcloud storage ls gs://my-bucket

Answer

gcloud storage buckets create gs://my-bucket

Answer

gsutil rsync -r /local/dir gs://my-bucket/

Question 14

A DevOps engineer needs to create a GKE cluster with 3 nodes per zone in us-central1-a and us-central1-b, with nodes of type e2-standard-2. The cluster should be regional. Which command is correct?

Accepted Answer

gcloud container clusters create my-cluster --region us-central1 --node-locations us-central1-a,us-central1-b --num-nodes 3 --machine-type e2-standard-2. To create a regional GKE cluster with 3 nodes per zone in us-central1-a and us-central1-b, the command must specify --region us-central1, --node-locations us-central1-a,us-central1-b, and --num-nodes 3. Option A includes all these parameters correctly.

Answer

gcloud container clusters create my-cluster --region us-central1 --machine-type e2-standard-2

Answer

gcloud container clusters create my-cluster --zone us-central1-a --region us-central1 --num-nodes 3 --machine-type e2-standard-2

Answer

gcloud container clusters create my-cluster --zone us-central1-a --num-nodes 3 --machine-type e2-standard-2

Question 15

A developer wants to deploy a containerized application to GKE and expose it externally via a load balancer. Which three kubectl commands are needed? (Choose THREE.)

Accepted Answer

kubectl create deployment my-app --image=gcr.io/my-project/my-app. Option B is correct because `kubectl create deployment my-app --image=gcr.io/my-project/my-app` creates the Deployment that runs the containerized application's Pods on GKE, which is the required first step before any Service can select and expose those Pods. Option C is correct because `kubectl expose deployment my-app --type=LoadBalancer --port=80 --target-port=8080` creates a Service of type LoadBalancer that provisions an external cloud load balancer, mapping external port 80 to the container's target port 8080. Option E is correct because `kubectl get service` is used to verify the Service was created and to retrieve its external IP/hostname once the load balancer is provisioned. Option A is not needed because `kubectl autoscale deployment` only configures a HorizontalPodAutoscaler and is unrelated to creating or exposing the workload. Option D is not needed because `kubectl run` creates a standalone Pod (or Deployment depending on version) and would duplicate the workload already created in option B rather than expose it.

Answer

kubectl autoscale deployment my-app --cpu-percent=50 --min=1 --max=10

Answer

kubectl run my-app --image=gcr.io/my-project/my-app