Associate-Cloud-Engineer Google Cloud Certified - Associate Cloud Engineer Questions and Answers

The correct answer is A because the application only needs to create new log objects in one specific bucket. A dedicated service account gives the workload its own identity, and granting roles/storage.objectCreator on the bucket limits access to the minimum scope. Google Cloud documents that Storage Object Creator allows object creation but does not grant permission to view, delete, or overwrite objects, and the role can be granted at the bucket level. Object Admin and Storage Admin are excessive because they allow broader object or bucket administration. Using the default Compute Engine service account at project scope is also too broad. Generating a service account key is unnecessary and weakens security.

This is the most optimal solution. Rather than recreating all nodes, you create a new node pool with GPU enabled. You then modify the pod specification to target particular GPU types by adding node selector to your workloads Pod specification. YOu still have a single cluster so you pay Kubernetes cluster management fee for just one cluster thus minimizing the cost.

Ref: https://cloud.google.com/kubernetes-engine/docs/how-to/gpus

Ref: https://cloud.google.com/kubernetes-engine/pricing

Example:

apiVersion: v1

kind: Pod

metadata:

name:my-gpu-pod

spec:

containers:

name:my-gpu-container

image: nvidia/cuda:10.0-runtime-ubuntu18.04

command: [/bin/bash]

resources:

limits:

nvidia.com/gpu: 2

nodeSelector:

cloud.google.com/gke-accelerator: nvidia-tesla-k80# or nvidia-tesla-p100 or nvidia-tesla-p4 or nvidia-tesla-v100 or nvidia-tesla-t4

https://cloud.google.com/bigquery/external-data-sources

An external data source is a data source that you can query directly from BigQuery, even though the data is not stored in BigQuery storage.

BigQuery supports the following external data sources:

Amazon S3

Azure Storage

Cloud Bigtable

Cloud Spanner

Cloud SQL

Cloud Storage

Drive

Comprehensive and Detailed Explanation From Exact Extract:

The key requirements are a containerized web application that handles unpredictable traffic and must minimize costs.

Cloud Run is the ideal solution because it is a fully managed, serverless platform for containers that automatically scales to zero instances when there is no traffic. This directly fulfills the cost requirement by eliminating charges for idle resources. It excels at handling unpredictable, bursty traffic.

GKE Standard and Autopilot (Options B and D) incur costs for the cluster or nodes even when not serving traffic (unless carefully scaled down), making them less cost-efficient than Cloud Run ' s native scale-to-zero for unpredictable, non-constant workloads.

Configure the ACLs on each image in Cloud Storage to give read-only access to the default Compute Engine service account. is not right.

As mentioned above, Container Registry ignores permissions set on individual objects within the storage bucket so this isnt going to work.

Ref: https://cloud.google.com/container-registry/docs/access-control

Adding an API as a type provider

This page describes how to add an API to Google Cloud Deployment Manager as a type provider. To learn more about types and type providers, read the Types overview documentation.

A type provider exposes all of the resources of a third-party API to Deployment Manager as base types that you can use in your configurations. These types must be directly served by a RESTful API that supports Create, Read, Update, and Delete (CRUD).

If you want to use an API that is not automatically provided by Google with Deployment Manager, you must add the API as a type provider.

https://cloud.google.com/deployment-manager/docs/configuration/type-providers/creating-type-provider

You can create two configurations – one for the development project and another for the production project. And you do that by running “gcloud config configurations create” command.

https://cloud.google.com/sdk/gcloud/reference/config/configurations/create

In your custom script, you can load these configurations one at a time and execute gcloud compute instances list to list Google Compute Engine instances in the project that is active in the gcloud configuration.

Ref: https://cloud.google.com/sdk/gcloud/reference/compute/instances/list

Once you have this information, you can export it in a suitable format to a suitable target e.g. export as CSV or export to Cloud Storage/BigQuery/SQL, etc

As mentioned in Schema and data model, you should be careful when choosing a primary key to not accidentally create hotspots in your database. One cause of hotspots is having a column whose value monotonically increases as the first key part, because this results in all inserts occurring at the end of your key space. This pattern is undesirable because Cloud Spanner divides data among servers by key ranges, which means all your inserts will be directed at a single server that will end up doing all the work.https://cloud.google.com/spanner/docs/schema-design#primary-key-prevent-hotspots

Implied rules Every VPC network has two implied firewall rules. These rules exist, but are not shown in the Cloud Console: Implied allow egress rule. An egress rule whose action is allow, destination is 0.0.0.0/0, and priority is the lowest possible (65535) lets any instance send traffic to any destination, except for traffic blocked by Google Cloud. A higher priority firewall rule may restrict outbound access. Internet access is allowed if no other firewall rules deny outbound traffic and if the instance has an external IP address or uses a Cloud NAT instance. For more information, see Internet access requirements. Implied deny ingress rule. An ingress rule whose action is deny, source is 0.0.0.0/0, and priority is the lowest possible (65535) protects all instances by blocking incoming connections to them. A higher priority rule might allow incoming access. The default network includes some additional rules that override this one, allowing certain types of incoming connections.https://cloud.google.com/vpc/docs/firewalls#default_firewall_rules

Navigate to Stackdriver Logging and select resource.labels.project_id=*. is not right.

Log entries are held in Stackdriver Logging for a limited time known as the retention period  which is 30 days (default configuration). After that, the entries are deleted. To keep log entries longer, you need to export them outside of Stackdriver Logging by configuring log sinks.

Ref: https://cloud.google.com/blog/products/gcp/best-practices-for-working-with-google-cloud-audit-logging

Configure a Cloud Scheduler job to read from Stackdriver and store the logs in BigQuery. Configure the table expiration to 60 days. is not right.

While this works, it makes no sense to use Cloud Scheduler job to read from Stackdriver and store the logs in BigQuery when Google provides a feature (export sinks) that does exactly the same thing and works out of the box.

Ref: https://cloud.google.com/logging/docs/export/configure_export_v2

Create a Stackdriver Logging Export with a Sink destination to Cloud Storage. Create a lifecycle rule to delete objects after 60 days. is not right.

You can export logs by creating one or more sinks that include a logs query and an export destination. Supported destinations for exported log entries are Cloud Storage, BigQuery, and Pub/Sub.

Ref: https://cloud.google.com/logging/docs/export/configure_export_v2

Sinks are limited to exporting log entries from the exact resource in which the sink was created: a Google Cloud project, organization, folder, or billing account. If it makes it easier to exporting from all projects of an organication, you can create an aggregated sink that can export log entries from all the projects, folders, and billing accounts of a Google Cloud organization.

Ref: https://cloud.google.com/logging/docs/export/aggregated_sinks

Either way, we now have the data in Cloud Storage, but querying logs information from Cloud Storage is harder than Querying information from BigQuery dataset. For this reason, we should prefer Big Query over Cloud Storage.

Create a Stackdriver Logging Export with a Sink destination to a BigQuery dataset. Configure the table expiration to 60 days. is the right answer.

You can export logs by creating one or more sinks that include a logs query and an export destination. Supported destinations for exported log entries are Cloud Storage, BigQuery, and Pub/Sub.

Ref: https://cloud.google.com/logging/docs/export/configure_export_v2

Sinks are limited to exporting log entries from the exact resource in which the sink was created: a Google Cloud project, organization, folder, or billing account. If it makes it easier to exporting from all projects of an organication, you can create an aggregated sink that can export log entries from all the projects, folders, and billing accounts of a Google Cloud organization.

Ref: https://cloud.google.com/logging/docs/export/aggregated_sinks

Either way, we now have the data in a BigQuery Dataset. Querying information from a Big Query dataset is easier and quicker than analyzing contents in Cloud Storage bucket. As our requirement is to Quickly analyze the log contents, we should prefer Big Query over Cloud Storage.

Also, You can control storage costs and optimize storage usage by setting the default table expiration for newly created tables in a dataset. If you set the property when the dataset is created, any table created in the dataset is deleted after the expiration period. If you set the property after the dataset is created, only new tables are deleted after the expiration period.

For example, if you set the default table expiration to 7 days, older data is automatically deleted after 1 week.

Ref: https://cloud.google.com/bigquery/docs/best-practices-storage

The correct answer is C because the goal is to test permissions as the service account without downloading or handling long-lived secrets. The roles/iam.serviceAccountTokenCreator role allows a user to impersonate a service account by generating short-lived credentials or access tokens. That is the recommended secure approach for local testing because it avoids service account key files, which are sensitive and difficult to rotate safely once distributed. The serviceAccountKeyAdmin role would allow key creation and management, which directly conflicts with the requirement to avoid secret handling. The workloadIdentityUser role is mainly for workload identity federation scenarios, not local user impersonation. The serviceAccountUser role allows attaching service accounts to resources but does not provide token creation for local testing.

auto-provisioning = Attaches and deletes node pools to cluster based on the requirements. Hence creating a GPU node pool, and auto-scaling would be betterhttps://cloud.google.com/kubernetes-engine/docs/how-to/node-auto-provisioning

The reason is that when you do health check, you want the VM to be working. Do the first check after initial setup time of 3 mins = 180 s < 200 s is reasonable.

The reason why our autoscaling is adding more instances than needed is that it checks 30 seconds after launching the instance and at this point, the instance isnt up and isnt ready to serve traffic. So our autoscaling policy starts another instance  again checks this after 30 seconds and the cycle repeats until it gets to the maximum instances or the instances launched earlier are healthy and start processing traffic  which happens after 180 seconds (3 minutes). This can be easily rectified by adjusting the initial delay to be higher than the time it takes for the instance to become available for processing traffic.So setting this to 200 ensures that it waits until the instance is up (around 180-second mark) and then starts forwarding traffic to this instance. Even after a cool out period, if the CPU utilization is still high, the autoscaler can again scale up but this scale-up is genuine and is based on the actual load.

Initial Delay Seconds  This setting delays autohealing from potentially prematurely recreating the instance if the instance is in the process of starting up. The initial delay timer starts when the currentAction of the instance is VERIFYING.

Ref: https://cloud.google.com/compute/docs/instance-groups/autohealing-instances-in-migs

The best option is to attach a single service account to the compute instances and add minimal rights to the service account. Then, allow the service account to impersonate a Cloud Identity user with elevated permissions to create, update, or delete resources. This way, the service account can use short-lived access tokens to authenticate to Google Cloud APIs without needing to manage service account keys. This option follows the principle of least privilege and reduces the risk of credential leakage and misuse.

Option A is not recommended because it requires human intervention, which can slow down the CI/CD pipeline and introduce human errors. Option C is not secure because it grants all required IAM permissions to a single service account, which can increase the impact of a compromised key. Option D is not cost-effective because it requires creating and managing multiple service accounts and keys, as well as using a secret manager service.

1: https://cloud.google.com/iam/docs/impersonating-service-accounts

2: https://cloud.google.com/iam/docs/best-practices-for-managing-service-account-keys

3: https://cloud.google.com/iam/docs/understanding-service-accounts

The most direct and recommended way to get a granular breakdown of your Google Cloud costs in BigQuery is to enable Cloud Billing data export to BigQuery when you create or manage your Cloud Billing account. This automatically sets up a daily export of your billing data to a BigQuery dataset you specify.

Option A: Enabling the BigQuery API and managing IAM roles are necessary for interacting with BigQuery, but they don ' t automatically populate it with Cloud Billing data. Selecting a data location is also important for BigQuery datasets but is a separate step from enabling billing export.

Option B: The BigQuery Data Transfer Service is used for transferring data from various sources into BigQuery, but for Cloud Billing data, the direct export feature is the standard and simpler method.

Option D: Enabling Cloud Billing and linking an account makes billing data available in the Cloud Billing console, but it doesn ' t automatically export it to BigQuery for detailed analysis. You need to explicitly configure the BigQuery export.

Reference to Google Cloud Certified - Associate Cloud Engineer Documents:

The process of setting up Cloud Billing export to BigQuery is clearly documented in the Google Cloud Billing documentation, which is a fundamental area for the Associate Cloud Engineer certification. Understanding how to access and analyze billing data is crucial for cost management.

Coldline Storage is a very-low-cost, highly durable storage service for storing infrequently accessed data. Coldline Storage is ideal for data you plan to read or modify at most once a quarter. Since we have a requirement to access data once a quarter and want to go with the most cost-efficient option, we should select Coldline Storage.

Ref: https://cloud.google.com/storage/docs/storage-classes#coldline

gcloud deployment-manager deployments create creates deployments based on the configuration file. (Infrastructure as code). All the configuration related to the artifacts is in the configuration file. This command correctly creates a cluster based on the provided cluster.yaml configuration file.

Ref: https://cloud.google.com/sdk/gcloud/reference/deployment-manager/deployments/create

The correct answer is B because independent applications should have independent Terraform root configurations and state boundaries. Google Cloud Terraform best practices state that to manage applications and projects independently, resources for each application and project should be placed in their own Terraform directories. This avoids a single oversized root module and reduces coupling between unrelated applications. Grouping by resource type creates artificial coupling because one application change can affect the same shared state as other applications. Putting all modules in one directory does not establish independent root configurations. A root directory per application gives clearer ownership, safer deployments, and better separation of lifecycle, state, and permissions.

https://cloud.google.com/resource-manager/docs/organization-policy/org-policy-constraints This list constraint defines the set of domains that email addresses added to Essential Contacts can have. By default, email addresses with any domain can be added to Essential Contacts. The allowed/denied list must specify one or more domains of the form @example.com. If this constraint is active and configured with allowed values, only email addresses with a suffix matching one of the entries from the list of allowed domains can be added in Essential Contacts. This constraint has no effect on updating or removing existing contacts. constraints/essentialcontacts.allowedContactDomains

This command correctly lists pods that have the label app=prod. When creating the deployment, we used the label app=prod so listing pods that have this label retrieve the pods belonging to nginx deployments. You can list pods by using Kubernetes CLI  kubectl get pods.

Ref: https://kubernetes.io/docs/tasks/access-application-cluster/list-all-running-container-images/

Ref: https://kubernetes.io/docs/tasks/access-application-cluster/list-all-running-container-images/#list-containers-filtering-by-pod-label

https://cloud.google.com/compute/docs/access/service-accounts#associating_a_service_account_to_an_instance

https://cloud.google.com/sdk/gcloud/reference/projects/create

Quickstart: Creating a New Instance Using the Command Line

Before you begin

1. In the Cloud Console, on the project selector page, select or create a Cloud project.

2. Make sure that billing is enabled for your Google Cloud project. Learn how to confirm billing is enabled for your project.

To use the gcloud command-line tool for this quickstart, you must first install and initialize the Cloud SDK:

1. Download and install the Cloud SDK using the instructions given on Installing Google Cloud SDK.

2. Initialize the SDK using the instructions given on Initializing Cloud SDK.

To use gcloud in Cloud Shell for this quickstart, first activate Cloud Shell using the instructions given on Starting Cloud Shell.

https://cloud.google.com/ai-platform/deep-learning-vm/docs/quickstart-cli#before-you-begin

Cloud Billing export to BigQuery enables you to export detailed Google Cloud billing data (such as usage, cost estimates, and pricing data) automatically throughout the day to a BigQuery dataset that you specify. Then you can access your Cloud Billing data from BigQuery for detailed analysis, or use a tool like Looker Studio to visualize your data.

The correct answer is B because programmatic budget notifications are intended for event-driven budget alert automation. Cloud Billing can publish budget threshold messages to Pub/Sub, and a Cloud Run function can be triggered by those messages to post to Slack. This gives near real-time notification without running persistent infrastructure. BigQuery billing export is valuable for analysis, but scheduled queries or DAGs are not near real time and add maintenance overhead. A Compute Engine managed instance group requires servers to run and manage, which is unnecessary for a simple notification workflow. Pub/Sub plus Cloud Run function is serverless, event-driven, easier to support, and directly aligned with Google Cloud’s cost-control automation pattern.

https://cloud.google.com/logging/docs/audit

Data Access audit logs Data Access audit logs contain API calls that read the configuration or metadata of resources, as well as user-driven API calls that create, modify, or read user-provided resource data.

https://cloud.google.com/logging/docs/audit#data-access

The correct answer is D because the team already uses GKE and wants the same declarative model and tooling used for application deployments. Config Connector lets teams manage Google Cloud resources as Kubernetes custom resources. This means infrastructure resources can be declared in Kubernetes manifests and applied through Kubernetes-native workflows. The Google Cloud console is manual and not declarative. Shell scripts with gcloud commands are imperative, harder to audit, and less aligned with Kubernetes deployment workflows. Terraform is declarative and valid for infrastructure management, but it uses separate tooling and state management. The question specifically asks for the same declarative approach and tooling as application deployments, making Config Connector the strongest match.

https://blog.doit-intl.com/the-truth-behind-google-cloud-egress-traffic-6e8f57b5c2f8

Keyword, Version, traffic splitting, App Engine supports traffic splitting for versions before releasing.

https://cloud.google.com/monitoring/alerts#alerting-example

https://cloud.google.com/kubernetes-engine/docs/how-to/cluster-autoscaler#adding_a_node_pool_with_autoscaling

A managed instance group (MIG) is a group of identical virtual machines (VMs) that you can manage as a single entity. You can use a MIG to deploy and maintain a stateless application that runs directly on VMs. A MIG can automatically scale the number of VMs based on the load or a schedule. A MIG can also automatically heal the VMs if they become unhealthy or unavailable. A MIG is suitable for applications that need to run on VMs rather than containers or serverless platforms.

B is incorrect because Kubernetes Engine is a managed service for running containerized applications on a cluster of nodes. It is not necessary to use Kubernetes Engine if the application does not use containers and can run directly on VMs.

C is incorrect because Cloud Functions is a serverless platform for running event-driven code in response to triggers. It is not suitable for applications that need to run continuously and handle HTTP requests.

D is incorrect because Cloud Run is a serverless platform for running stateless containerized applications. It is not suitable for applications that do not use containers and can run directly on VMs.

Managed instance groups documentation

Choosing a compute option for Google Cloud

According to the Google Cloud documentation, you can use organization policy constraints to control the creation and expiration of service account keys. The constraints are:

constraints/iam.allowServiceAccountKeyCreation: This constraint allows you to specify which projects or folders can create service account keys. You can set the value to true or false, or use a condition to apply the constraint to specific service accounts. By setting this constraint to false for the organization and adding an exception for the pj-sa project, you can prevent developers from creating service account keys in other projects.

constraints/iam.serviceAccountKeyMaxLifetime: This constraint allows you to specify the maximum lifetime of service account keys. You can set the value to a duration in seconds, such as 86400 for one day. By setting this constraint to 86400 for the organization, you can ensure that all service account keys expire after one day.

These constraints are recommended by Google Cloud as best practices to minimize the risk of service account key misuse or compromise. They also help you reduce the cost of managing service account keys, as you do not need to implement a custom solution to rotate or delete them.

1: Associate Cloud Engineer Certification Exam Guide | Learn - Google Cloud

5: Create and delete service account keys - Google Cloud

Organization policy constraints for service accounts

The requirement is for an infrastructure-as-code (IaC) solution that can manage Kubernetes resources and other cloud resources across multiple cloud providers (Google Cloud and two others).

Option A (YAML manifests): YAML manifests are primarily used for defining Kubernetes objects, not for provisioning general cloud resources (like VPCs, IAM policies, databases) across different cloud providers.

Option C (Python + SDKs): While possible, writing custom scripts using each provider ' s SDK requires significant development effort to handle state management, dependencies, and provider differences. It essentially reinvents much of what dedicated IaC tools provide and is not a standard IaC approach.

Option D (Config Connector): Config Connector allows managing Google Cloud resources using Kubernetes-style manifests and APIs. It is specific to Google Cloud and cannot manage resources in other cloud providers.

Option B (Terraform): Terraform is an open-source IaC tool explicitly designed for building, changing, and versioning infrastructure safely and efficiently across multiple cloud providers and on-premises data centers. It uses providers for different platforms (GCP, AWS, Azure, Kubernetes, etc.), allowing a unified workflow to manage diverse resources across the required environments (Google Cloud, other clouds, Kubernetes).

Terraform is the standard tool for multi-cloud IaC automation as described in the scenario.

Keywords, Financial data (large data) used globally, data stored and queried using relational structure (SQL), clients should get exact identical copies(Strong Consistency), Multiple region, low latency to end user, select storage option to minimize latency.

When encountering a quotaExceeded error in BigQuery, you should follow these steps to diagnose and mitigate the issue:

Understand the Error:

The error message indicates that a quota was exceeded (either a short-term rate limit or a longer-term limit).

The response payload contains information about which quota was reached.

Quotas can fall into two categories:

rateLimitExceeded: Short-term limits. Retry the operation after a few seconds using exponential backoff.

quotaExceeded: Longer-term limits. Wait 10 minutes or longer before retrying the operation.

Search Errors in Cloud Audit Logs (Option A):

Cloud Audit Logs provide detailed information about API requests and responses.

By searching the logs, you can identify the specific API call that triggered the quotaExceeded error.

This helps you understand which resource or operation exceeded the quota.

View Errors in Cloud Monitoring (Option C):

Cloud Monitoring (formerly known as Stackdriver) provides insights into your Google Cloud resources.

Check the monitoring dashboard for any alerts related to BigQuery quotas.

You can set up custom monitoring rules to track specific quotas and receive notifications.

Other Options:

B. Configure Cloud Trace: Cloud Trace is used for performance analysis and latency tracking. It’s not directly related to quota issues.

D. Use Information Schema Views: Information schema views provide metadata about your datasets and tables but won’t help diagnose quota errors.

E. Use BigQuery Bl Engine: There is no such tool called “BigQuery Bl Engine.” This option is invalid.

Remember that some quotas replenish incrementally over a 24-hour period, so you don’t always need to wait a full 24 hours after reaching the limit. Ifyou consistently hit longer-term quotas, consider workload optimization or requesting a quota increase

Comprehensive and Detailed Explanation From Exact Extract:

The issue is Cloud SQL API quota errors caused by too many new Cloud Run instances being created simultaneously during a traffic spike. Each new instance attempts a connection to Cloud SQL, consuming the connection API quota.

The Documented Solution (Not an Option): Google ' s official documentation states the correct mitigation for this specific error is to decrease the maximum number of Cloud Run instances to slow the rate of new instance creation.

Flawed Option Set: Since the correct mitigation is not an option, and increasing the maximum instances (D) would technically worsen the quota error, the question is structurally flawed. However, in the context of general autoscaling management, Option D is the configuration control point for the upper limit of scaling. In many exam scenarios, this forces the selection of the control that directly impacts the scaling capacity, even if the intended direction is incorrect for the specific error given. We select D as the configuration change most directly related to the number of instances, while acknowledging the documentation advises the opposite action.

Keywords: - continually - > Standard - mission-critical analytics - > dual-regional

The correct answer is A because Hyperdisk Extreme is designed for workloads that require very high IOPS and low-latency block storage, which fits SAP HANA performance requirements. The question also explicitly requires the data to persist through instance restarts. Hyperdisk Extreme is persistent block storage, so the application data remains available after the VM restarts. In-memory RAM disks are not suitable because data is volatile and lost when the instance restarts. Local SSD can provide high performance, but it is ephemeral and does not preserve data across stop/start or certain maintenance events, so it fails the persistence requirement. Hyperdisk Throughput is optimized for throughput-heavy workloads, not the highest IOPS profile required here.

Creating or upgrading a cluster by specifying the version as latest does not provide automatic upgrades. Enable node auto-upgrades to ensure that the nodes in your cluster are up-to-date with the latest stable version.

https://cloud.google.com/kubernetes-engine/versioning-and-upgrades

Node auto-upgrades help you keep the nodes in your cluster up to date with the cluster master version when your master is updated on your behalf. When you create a new cluster or node pool with Google Cloud Console or the gcloud command, node auto-upgrade is enabled by default.

Ref: https://cloud.google.com/kubernetes-engine/docs/how-to/node-auto-upgrades

The correct answer is A because the requirement is company-wide access for employees in one Cloud Identity domain, not public or cross-domain access. The BigQuery Data Viewer role provides read-only access to dataset resources, including reading and querying table data when granted at the dataset level. Using the domain principal domain:terramearth.com grants access to authenticated users in that organization’s domain. allUsers would include unauthenticated public access, and allAuthenticatedUsers would include any authenticated Google account, which is far broader than the company boundary. A service account represents an application identity, not all human employees.

The correct answer is D because Google-recommended IAM administration favors group-based access rather than assigning roles directly to individual users. Cloud Identity groups make onboarding and offboarding simpler: when a person joins or leaves a team, administrators update group membership instead of editing IAM bindings across hundreds of resources. Security groups are specifically intended for access-control use cases and are better suited for IAM policies than ordinary discussion-only groups. Service accounts should represent workloads, not teams of human users. Granting humans permission to impersonate shared service accounts weakens accountability, and allowing users to generate service account keys is even worse because it introduces long-lived credentials. Security-labeled Cloud Identity groups provide scalable, auditable access management.

https://cloud.google.com/iap/docs/using-tcp-forwarding

https://cloud.google.com/sdk/gcloud/reference/compute/networks/subnets/expand-ip-range

gcloud compute networks subnets expand-ip-range - expand the IP range of a Compute Engine subnetwork gcloud compute networks subnets expand-ip-range NAME --prefix-length=PREFIX_LENGTH [--region=REGION] [GCLOUD_WIDE_FLAG …]

To minimize latency for global users and support growth across multiple geographical regions, you must deploy your resources in those regions and use a global load balancing service.

Action B (Multi-Region Deployment): Deploying VMs in multiple regions closest to your users is necessary to reduce latency (by shortening the physical distance data must travel) and support regional growth.

Action C (Global External Application Load Balancer): The external Application Load Balancer (Global) is the Google-recommended Layer 7 solution for applications that require global reach. It provides a single anycast IP address and directs users to the nearest healthy backend, thus minimizing latency globally.

The correct answer is C because Cloud Asset Inventory can search resources across an organization and filter by asset type and creation time. The asset type iam.googleapis.com/ServiceAccountKey directly targets service account keys, which is exactly what must be audited. Using organization scope is important because the question says there are hundreds of service accounts across applications, likely distributed across projects. Cloud KMS keys are encryption keys, not service account keys, so option A searches the wrong resource type. Listing service accounts only identifies service accounts, not their individual keys or key ages. API keys are different credentials from service account keys, so option D also targets the wrong asset type. Cloud Asset Inventory is the most scalable audit method here.

App engine is a regional service, which means the infrastructure that runs your app(s) is located in a specific region and is managed by Google to be redundantly available across all the zones within that region. Once an app engine deployment is created in a region, it cant be changed. The only way is to create a new project and create an App Engine instance in europe-west3, send all user traffic to this instance and delete the app engine instance in us-central.

Ref: https://cloud.google.com/appengine/docs/locations

The correct answer is B because the Organization Policy Service provides a managed resource locations constraint for restricting where supported location-based resources can be created. Using the gcp.resourceLocations list constraint with an allowed_values entry is the cleanest and lowest-maintenance approach. The in:australia-locations value represents Google’s predefined location group for Australia locations, avoiding manual custom logic. A denied_values policy would block only specified locations and could accidentally allow other regions, which does not satisfy the strict regulatory requirement. Custom CEL policies add unnecessary complexity and are less appropriate when a built-in list constraint exists for this exact purpose. The requirement asks for minimal configuration, so using the predefined location group in allowed_values is the strongest fit.

(https://cloud.google.com/resource-manager/docs/project-migration#change_billing_account)

https://cloud.google.com/billing/docs/concepts

https://cloud.google.com/resource-manager/docs/project-migration

Google Cloud Coldline is a new cold-tier storage for archival data with access frequency of less than once per year. Unlike other cold storage options, Nearline has no delays prior to data access, so now it is the leading solution among competitors.

The Real description is about Coldline storage Class:

Coldline Storage

Coldline Storage is a very-low-cost, highly durable storage service for storing infrequently accessed data. Coldline Storage is a better choice than Standard Storage or Nearline Storage in scenarios where slightly lower availability, a 90-day minimum storage duration, and higher costs for data access are acceptable trade-offs for lowered at-rest storage costs.

Coldline Storage is ideal for data you plan to read or modify at most once a quarter. Note, however, that for data being kept entirely for backup or archiving purposes, Archive Storage is more cost-effective, as it offers the lowest storage costs.

https://cloud.google.com/storage/docs/storage-classes#coldline

For a web application (typically using HTTP/HTTPS), an Application Load Balancer is the recommended choice as it operates at Layer 7, providing features like content-based routing, SSL termination, and improved security. To expose the application publicly, you would need to use a public DNS zone. An A record in a public DNS zone maps a domain name to the public IP address of the Application Load Balancer. Using a CNAME record would also work but is generally recommended for aliasing one domain name to another, not directly to an IP address.

Option A & B: Network Load Balancers operate at Layer 4 (TCP/UDP) and lack the application-level features of an Application Load Balancer. Private DNS zones are for internal name resolution within your VPC, not for public access.

Option C: While an Application Load Balancer is the correct type, using a private DNS zone wouldn ' t make the web application publicly accessible.

Reference to Google Cloud Certified - Associate Cloud Engineer Documents:

The best practices for load balancing web applications on Google Cloud, including the use of Application Load Balancers for Layer 7 traffic and the configuration of public DNS records (A or CNAME) for public access, are detailed in the Google Cloud Load Balancing and Cloud DNS documentation, both important for the Associate Cloud Engineer certification.

Cloud Storage allows world-wide storage and retrieval of any amount of data at any time. We dont need to set up auto-scaling ourselves. Cloud Storage autoscaling is managed by GCP. Cloud Storage is an object store so it is suitable for storing photos. Cloud Storage allows world-wide storage and retrieval so cater well to our worldwide audience. Cloud storage provides us lifecycle rules that can be configured to automatically delete objects older than 30 days. This also fits our requirements. Finally, Google Cloud Storage offers several storage classes such as Nearline Storage ($0.01 per GB per Month) Coldline Storage ($0.007 per GB per Month) and Archive Storage ($0.004 per GB per month) which are significantly cheaper than any of the options above.

Ref: https://cloud.google.com/storage/docs

Ref: https://cloud.google.com/storage/pricing

https://cloud.google.com/spanner/docs/getting-started/set-up

Billing Administrators can not create a new billing account, and the project is presumably already created. Project Billing Manager allows you to link the created billing account to the project. It is vague on how the billing account gets created but by process of elimination

The requirements specify a relational database that provides global scale and is a managed service.

Cloud Spanner is Google Cloud ' s globally distributed, strong-consistency database service. It provides limitless horizontal scaling while maintaining the structure of a relational database, making it the best fit for an application with a rapidly growing, global user base. As a fully managed service, it inherently minimizes management and administration efforts.

Cloud SQL (Option A) is a regional managed relational database and does not offer the same unlimited, global horizontal scaling capability as Spanner.

https://cloud.google.com/vpc/docs/vpc-peering

From this source:https://cloud.google.com/billing/docs/how-to/custom-roles#permission_association_and_inheritance

" For example, associating a project with a billing account requires the billing.resourceAssociations.create permission on the billing account and also the resourcemanager.projects.createBillingAssignment permission on the project. This is because project permissions are required for actions where project owners control access, while billing account permissions are required for actions where billing account administrators control access. When both should be involved, both permissions are necessary. "

https://gtseres.medium.com/using-service-accounts-across-projects-in-gcp-cf9473fef8f0#:~:text=Go%20to%20the%20destination%20project,Voila!

In Google compute engine, if predefined machine types don’t meet your needs, you can create an instance with custom virtualized hardware settings. Specifically, you can create an instance with a custom number of vCPUs and custom memory, effectively using a custom machine type. Custom machine types are ideal for the following scenarios:

   1. Workloads that aren’t a good fit for the predefined machine types that are available to you.

   2. Workloads that require more processing power or more memory but don’t need all of the upgrades that are provided by the next machine type level.

In our scenario, we only need a memory upgrade. Moving to a bigger instance would also bump up the CPU which we don’t need so we have to use a custom machine type. It is not possible to change memory while the instance is running so you need to first stop the instance, change the memory and then start it again. See below a screenshot that shows how CPU/Memory can be customized for an instance that has been stopped.

Ref: https://cloud.google.com/compute/docs/instances/creating-instance-with-custom-machine-type

Comprehensive and Detailed Explanation From Exact Extract:

The requirements point to the highest performance tier for a managed PostgreSQL-compatible database, specifically citing demands for transactional and analytical performance, synchronous replication, and storage optimization.

AlloyDB for PostgreSQL is Google Cloud ' s highly specialized, fully managed database that is optimized for demanding enterprise workloads. It is superior to standard Cloud SQL for performance and architectural features like synchronous replication and an optimized storage layer, directly matching all the stated requirements.

Database Migration Service (DMS) is the recommended tool for performing the migration itself.

In general, Google recommends that each instance that needs to call a Google API should run as a service account with the minimum permissions necessary for that instance to do its job. In practice, this means you should configure service accounts for your instances with the following process: Create a new service account rather than using the Compute Engine default service account. Grant IAM roles to that service account for only the resources that it needs. Configure the instance to run as that service account. Grant the instance the https://www.googleapis.com/auth/cloud-platform scope to allow full access to all Google Cloud APIs, so that the IAM permissions of the instance are completely determined by the IAM roles of the service account. Avoid granting more access than necessary and regularly check your service account permissions to make sure they are up-to-date.https://cloud.google.com/compute/docs/access/create-enable-service-accounts-for-instances#best_practices

The correct answer is A because the policy must apply across projects that are already spread across different folders. Setting the resource location constraint at the organization level gives consistent inheritance across the whole resource hierarchy instead of requiring repeated folder-level configuration. Since the business requirement is to permit only EEA locations, an allow-list model is safer than a deny-list model. A deny list can miss future or unexpected locations, while allowed values define the permitted geographic boundary. Google Cloud’s resource locations constraint supports allowed_values and denied_values, and Google recommends using allowed values or value groups to define the geographic boundary for resources.

The goal is to pull (fetch) messages from a subscription and acknowledge them.

The gcloud pubsub subscriptions **pull** command retrieves messages from a specified subscription.

The --auto-ack flag instructs the command to automatically acknowledge the messages after they are successfully retrieved, combining the two required actions into one command.

The correct answer is B because Cloud Storage Object Lifecycle Management is designed to automate storage-class transitions based on object conditions such as object age. The requirement is not to delete the records, but to retain them indefinitely while reducing cost after 90 days. Archive storage is the lowest-cost class for long-term retained data that is rarely accessed. The SetStorageClass lifecycle action changes the storage class of an object when the rule conditions are met. Object Versioning and daysSinceNoncurrentTime apply to noncurrent object versions, which is not the scenario. A scheduled Cloud Run function would add unnecessary operational overhead. A Delete action would violate the retention requirement.

Shared VPC allows an organization to connect resources from multiple projects to a common Virtual Private Cloud (VPC) network, so that they can communicate with each other securely and efficiently using internal IPs from that network. When you use Shared VPC, you designate a project as a host project and attach one or more other service projects to it. The VPC networks in the host project are called Shared VPC networks. Eligible resources from service projects can use subnets in the Shared VPC network

https://cloud.google.com/vpc/docs/shared-vpc

" For example, an existing instance in a service project cannot be reconfigured to use a Shared VPC network, but a new instance can be created to use available subnets in a Shared VPC network. "

When setting support levels for permissions in custom roles, you can set to one of SUPPORTED, TESTING or NOT_SUPPORTED.

Ref: https://cloud.google.com/iam/docs/custom-roles-permissions-support

The correct answer is B because Cloud Storage is the appropriate Google-managed object storage service for durable document storage, and Object Lifecycle Management can automatically transition objects to a lower-cost storage class after a defined age. The documents are frequently accessed for the first six months, so keeping them initially in Standard storage is reasonable. After six months, the documents are retained mainly for audit purposes, making Coldline more cost-effective for infrequently accessed data. Option A is backward because Standard is the higher-access storage class, not the archive target. Filestore is a managed file system, not the best low-cost object archive for invoices. Moving data manually or through Filestore adds unnecessary operational overhead.

https://cloud.google.com/storage/docs/lifecycle#setstorageclass-cost

# The object ' s time spent set at the original storage class counts towards any minimum storage duration that applies for the new storage class.

The requirement is for private communication between a Cloud Run service and a GKE API, following best practices.

Option A exposes the GKE API to the public internet, which violates the " privately reach " requirement. Relying on dynamic IP allowlisting with Cloud Armor is complex and less secure than private networking.

Options B and C configure overly permissive firewall rules (allowing all egress or ingress) and do not establish the necessary private network path between Cloud Run (which normally runs outside your VPC) and the GKE cluster within your VPC.

Option D describes the standard Google-recommended pattern for this scenario:

Internal Application Load Balancer (ILB): Expose the GKE service (API) using an ILB. This gives the service a private IP address accessible only within the VPC network (or connected networks).

Cloud DNS: Create a private DNS zone and record pointing a fully qualified domain name (FQDN) to the ILB ' s private IP address. This allows services to reach the API via a stable name instead of an IP.

Serverless VPC Access Connector: This connector creates a bridge allowing serverless services like Cloud Run to send traffic into your VPC network.

Cloud Run Configuration: Configure the Cloud Run service to use the VPC Access connector. The application code can then call the GKE API using its private FQDN registered in Cloud DNS.

This setup ensures traffic flows entirely over private networks (within the VPC via the ILB and through the VPC Access connector), meeting the private communication requirement securely and reliably.

https://cloud.google.com/monitoring/settings/multiple-projects

https://cloud.google.com/monitoring/workspaces

This aligns with Googles recommended practices. By creating a new project, we achieve complete isolation between development and production environments; as well as isolate this production application from production applications of other departments.

Ref: https://cloud.google.com/docs/enterprise/best-practices-for-enterprise-organizations#define-hierarchy

`gcloud services list --available` returns not only the enabled services in the project but also services that CAN be enabled.

https://cloud.google.com/sdk/gcloud/reference/services/list#--available

Run the following command to list the enabled APIs and services in your current project:

gcloud services list

whereas, Run the following command to list the APIs and services available to you in your current project:

gcloud services list –available

https://cloud.google.com/sdk/gcloud/reference/services/list#--available

--available

Return the services available to the project to enable. This list will include any services that the project has already enabled.

To list the services the current project has enabled for consumption, run:

gcloud services list --enabled

To list the services the current project can enable for consumption, run:

gcloud services list –available

The correct answer is D because Kubernetes provides a direct rollback command for failed or problematic rollouts. In GKE, deployments are managed using standard Kubernetes deployment mechanics, so kubectl rollout undo is the correct command to return a workload to the previous rollout revision. Kubernetes documentation describes kubectl rollout undo as the command used to roll back to a previous rollout and gives deployment rollback examples. Manually scaling old and new Pods is risky because it bypasses deployment controller state and can create configuration drift. Deleting the deployment causes service disruption and removes the workload object. kubectl rollout restart restarts Pods using the current deployment specification; it does not revert to a previous revision.

https://cloud.google.com/architecture/best-practices-vpc-design#shared-service Cloud VPN is another alternative. Because Cloud VPN establishes reachability through managed IPsec tunnels, it doesn ' t have the aggregate limits of VPC Network Peering. Cloud VPN uses a VPN Gateway for connectivity and doesn ' t consider the aggregate resource use of the IPsec peer. The drawbacks of Cloud VPN include increased costs (VPN tunnels and traffic egress), management overhead required to maintain tunnels, and the performance overhead of IPsec.

A deployment is responsible for keeping a set of pods running. A service is responsible for enabling network access to a set of pods.

https://developer.mozilla.org/en-US/docs/Web/HTTP/Basics_of_HTTP/MIME_Types#importance_of_setting_the_correct_mime_type

gcloud compute ssh ensures that the user’s public SSH key is present in the project’s metadata. If the user does not have a public SSH key, one is generated using ssh-keygen and added to the project’s metadata. This is similar to the other option where we copy the key explicitly to the project’s metadata but here it is done automatically for us. There are also security benefits with this approach. When we use gcloud compute ssh to connect to Linux instances, we are adding a layer of security by storing your host keys as guest attributes. Storing SSH host keys as guest attributes improve the security of your connections by helping to protect against vulnerabilities such as man-in-the-middle (MITM) attacks. On the initial boot of a VM instance, if guest attributes are enabled, Compute Engine stores your generated host keys as guest attributes.

Compute Engine then uses these host keys that were stored during the initial boot to verify all subsequent connections to the VM instance.

Ref: https://cloud.google.com/compute/docs/instances/connecting-to-instance

Ref: https://cloud.google.com/sdk/gcloud/reference/compute/ssh