Prepare and test your skills

Prepare and test your skills

Utilize Verification Tools for Service Availability means confirming that a specific Google Cloud (opens in a new tab) service, feature, or capacity exists in a target region or zone before deployment. The three primary verification tools are the Google Cloud Console for visual checks, the gcloud command-line interface (gcloud CLI) for terminal queries, and the official Cloud Locations documentation for product-by-location lists. Verification must account for resource scope, because global, regional, and zonal resources follow different rules for where they can be used.
Resource scope in Google Cloud defines which resources can reach each other across locations. A global resource is accessible from any region or zone in the project, a regional resource is accessible only within the same region, and a zonal resource is accessible only within the same zone. Zonal resources can still use regional and global resources in their location, but the reverse is not true across boundaries. The scope of an operation follows the scope of its resource, so creating a Virtual Private Cloud (VPC) network is a global operation while reserving a static external IP address is a regional operation.
| Scope | Accessibility | Naming uniqueness | Examples |
|---|---|---|---|
| Global | Any region or zone in the project | Project-wide | Images, VPC networks, global instance templates |
| Regional | Any zone within the same region | Region-wide | Subnets, regional static IPs, regional disks |
| Zonal | Same zone only | Zone-wide | Compute Engine VM instances, zonal disks, machine types |
A zone name combines its region with a letter, such as asia-east1-a for zone a in region asia-east1. Zones in the same region have high-bandwidth, low-latency connections, but communication inside one region is generally cheaper and faster than communication across regions. Within a project, a zonal name must be unique only in its zone, so two virtual machines (VMs) named my-instance can exist in us-central1-a and us-central1-b.
Exam tip: Disks can only be attached to computers in the same zone, so verify both the VM and the disk in that exact zone.
Verification in the Google Cloud Console uses visual pages to confirm regions, zones, capacity, quotas, and service health. The console can list available regions and zones for Compute Engine, and during instance creation you can select Any for the zone to let Google choose a zone based on machine type and availability. The Quotas page shows project-wide and per-region limits, so a blocked creation may mean a quota was reached rather than a service being unavailable in that location.
The console also checks current capacity and health for a target location. The Capacity advisor page checks Spot VM availability for selected regions, machine families, series, types, and VM count:
The Google Cloud Service Health Dashboard shows current status of Google Cloud services in every location, with incident history including duration, affected zones and regions, and workarounds. Personalized Service Health shows incidents relevant to a specific project and supports per-project or per-organization API access and alerts.
Verification with the gcloud CLI uses query commands to confirm locations, machine types, and capacity from a terminal. After installing the gcloud CLI and running gcloud init, you can set a default region and zone for the local client. Running gcloud components update ensures the CLI has the latest location data and commands.
Specific commands return availability for a service or configuration in target locations. The command gcloud compute machine-types list --filter="name=c4d-standard-4" lists all regions and zones where that machine type can be used. The command gcloud privateca locations list lists supported locations for Certificate Authority Service resources such as certificates, certificate authorities, and CA pools. The command gcloud beta compute advice capacity with flags for provisioning model, machine types, target distribution shape, size, and region checks Spot VM availability for regional managed instance group planning.
Verification with the official Cloud Locations documentation confirms whether Google offers a product as global, multi-region, regional, or zonal. The Cloud Locations page at cloud.google.com/about/locations includes a Products available by location list and a Global products list for services available across all locations. Multi-region resources such as dual-region and multi-region Cloud Storage buckets, multi-region Spanner instances, and multi-region key rings are listed separately from global resources such as projects, VPC networks with routes and firewall rules, Cloud DNS zones, global external Application Load Balancers, Pub/Sub topics, and Secret Manager secrets.
The Geography and regions documentation explains region-specific considerations that affect placement, latency, and failure isolation. Regions are independent geographic areas made of zones, and distributing resources across zones protects against power, cooling, and networking failures while distributing across regions protects against larger outages. As a design rule, keep application components geographically near each other and near clients, because a CA pool cannot be moved after creation and cross-region attachment is restricted for performance and reliability.
Differentiate Service Scopes and Resource Hierarchy means understanding where global, regional, and zonal resources (opens in a new tab) can be used and what failures they can survive. The four scopes are zonal, regional, multi-region, and global, built on zones grouped into regions. The resource hierarchy places these resources inside projects, and projects themselves are global resources that can hold resources in any location.
Zonal resources exist in a single zone and can only be used by resources in that same zone. A Compute Engine VM instance is a zonal resource, along with its machine type, zonal Persistent Disks, Google Cloud Hyperdisk volumes, single-zone Google Kubernetes Engine (GKE) clusters, Filestore Basic and Zonal instances, Dataflow jobs, and Cloud SQL instances. A VM in one zone can still use regional and global resources, such as a regional subnet or a global image, but a zonal disk can only be attached to a VM in the same zone.
A zone defines the expected boundary for a failure. A failure in one zone usually does not affect resources in other zones in the same region, and zones are designed to minimize correlated failures. To protect zonal resources, deploy them redundantly across multiple zones or regions, for example by running VMs in a regional managed instance group (MIG) instead of as standalone VMs. Single-zone deployments target 99.9% availability, or about 43.2 minutes of downtime in a 30-day month.
Regional resources are deployed redundantly across multiple zones within the same region and can be used by any resource in that region. Regional resources survive a single-zone incident because data and capacity exist in other zones in the same region. A region outage from a natural disaster or large-scale infrastructure failure can still affect some or all regional resources in that region.
Examples of regional resources include:
Deployments across multiple zones in a region target 99.99% availability, or about 4.3 minutes of downtime in a 30-day month, so for higher protection deploy redundantly in multiple regions.
Multi-region resources are distributed across specific groups of regions for resilience to regional incidents. A multi-region resource remains available when one of its regions or zones has an outage, unless multiple regions fail at the same time. Deployments across multiple regions target 99.999% availability, or about 26 seconds of downtime in a 30-day month.
Common multi-region options show how the same product can have different scopes:
Choose a multi-region configuration when data must survive a regional outage without manual failover, and choose a regional configuration when data must stay within one region for latency, cost, or regulatory reasons.
Global resources are available across all Google Cloud locations and do not depend on infrastructure in any specific region or zone. Global resources are resilient to zonal and regional incidents. Because their configuration is globally scoped, one global resource is easier to operate than many regional copies, but a critical configuration error can become a single point of failure across all geographies.
Examples of global resources include:
Manage changes to global resources carefully, and for very high availability needs use regional defense-in-depth fallbacks, such as regional load balancers behind a global load balancer with automated failover using Cloud DNS.
Exam tip: Global resources survive zone and region outages, but a misconfiguration of a single global load balancer can affect all geographies at once.
Access rules flow from narrow to wide across scopes. Zonal resources can use zonal, regional, and global resources in their location, regional resources can use other resources in the same region, and global resources can be reached from any region or zone in the project. Naming follows the same boundary, with zonal names unique within the zone, regional names unique within the region, and global names unique project-wide.
| Resource scope | Resilience | Mitigation for infrastructure outages |
|---|---|---|
| Zonal | Low | Deploy redundantly in multiple zones or regions |
| Regional | Medium | Deploy redundantly in multiple regions |
| Multi-region or global | High | Manage changes carefully and use defense-in-depth fallbacks |
A combined architecture shows how the scopes work together. Traffic flows from clients to a global external Application Load Balancer, then from the load balancer to a regional MIG backend with Compute Engine VMs in two zones. Data flows from the application on the VMs to a Cloud SQL primary instance, then by synchronous replication to a standby instance in another zone using Hyperdisk Balanced High Availability volumes. Backups flow from the database to a multi-region Cloud Storage bucket, so a zone failure triggers automatic failover to the standby while a region failure allows restore in a different region from the backup.
Evaluate Location-Based Constraints and Requirements means checking whether a chosen Google Cloud location meets compliance, performance, and availability needs before deploying resources. A location decision controls where data is stored and accessed, how far traffic must travel between services, and which redundancy options are possible. Choosing the wrong region can create regulatory risk, extra egress cost and latency, or lower uptime.
Data residency compliance means keeping regulated data inside an approved geographic boundary and controlling who can access it from outside that boundary. Sovereign Controls by Partners (opens in a new tab) supports this boundary with technical controls that shape where data flows and how access is approved.
Sovereign Controls by Partners provides:
Even with these controls, common metadata fields may not be protected by the same technical boundary. Avoid placing sensitive or regulated data in resource names and IDs, project or folder names and IDs, description fields or labels, log-based metrics, VM sizes and similar service configurations, URIs or file paths, timestamps, user IDs, firewall rules, security scanning configurations, and customer Identity and Access Management (IAM) policies. Keeping sensitive values out of these fields helps maintain compliance because names, labels, descriptions, and configurations can be visible outside the protected data boundary.
Network latency and egress cost increase when resources that work together are placed in different regions. Traffic flows from the compute resource in one region to the storage resource in another region, so every read or write pays a distance penalty in time and in cross-region egress charges. For example, a Compute Engine instance in us-central1 that frequently accesses a Cloud Storage bucket in us-east1 creates a sustained cross-region flow that raises cost and slows responses.
The fix is to colocate dependent resources in the same region. Cloud Storage supports this through bucket relocation, which moves the bucket to the same region as the Google Cloud services that use it. Colocation keeps traffic inside one region, which reduces latency and avoids unexpected cross-region egress.
Exam tip: Colocate Compute Engine and Cloud Storage in the same region; a Compute Engine instance in one region reading a bucket in another region causes cross-region egress.
Availability design means selecting a topology that keeps connectivity working during maintenance or hardware failure. Partner Cross-Cloud Interconnect for OCI offers two prescriptive configurations that use the Cloud Interconnect service level agreement (SLA) for the Google Cloud portion of the path, up until responsibility transfers to OCI.
The two availability options differ by tolerance for downtime and deployment scope:
| Configuration | Choose when | Deployment scope |
|---|---|---|
| 99.99% availability | The application is production-level with low tolerance for downtime | Multi-region deployment, which provides higher availability than a single region |
| 99.9% availability | The application is not mission-critical and can tolerate some downtime | Single-region or simpler deployment with less redundancy |
The 99.99% availability topology requires at least four VLAN attachments across two metros, with one attachment in each edge availability domain. The design also requires two Cloud Routers in two Google Cloud regions of a VPC network, with one Cloud Router associated with each pair of VLAN attachments. The VPC network must enable global routing, and an equal number of OCI FastConnect virtual circuits must be created in the paired OCI location. When traffic leaves the VPC network, Google Cloud uses equal-cost multi-path routing (ECMP) to balance egress across the VLAN attachments based on configured capacity, which requires the Cloud Routers to receive the same announcement with the same CIDR range and the same MED values.
An isolation boundary means defining a perimeter that keeps data and API calls inside approved resources and networks. VPC Service Controls (VPC-SC) creates isolation perimeters around Google Cloud managed services, resources, and networks to help mitigate data exfiltration. Traffic flows from clients and workloads inside the perimeter to protected services inside the same perimeter, while requests to resources outside the perimeter are denied by policy.
Perimeter protection applies at the project level for some services. Storage Intelligence advisor can be enabled at the project, folder, or organization level, but VPC-SC helps protect only project-level resources because VPC-SC does not support adding folder-level or organization-level resources to service perimeters. Folder-level and organization-level oversight for Storage Intelligence advisor depends on IAM instead.
A global resource is accessible from any region or zone in the project, a regional resource is accessible only within the same region, and a zonal resource is accessible only within the same zone. Zonal resources can still use regional and global resources in their location, but the reverse is not true across boundaries, and naming uniqueness follows the same boundary with zonal names unique within the zone, regional names unique within the region, and global names unique project-wide.
The three primary verification tools are the Google Cloud Console for visual checks, the gcloud command-line interface (gcloud CLI) for terminal queries, and the official Cloud Locations documentation for product-by-location lists. The Google Cloud Console can list available regions and zones for Compute Engine and show quotas, capacity, and health, while the Cloud Locations page at cloud.google.com/about/locations includes a Products available by location list and a Global products list.
Dependent Compute Engine and Cloud Storage resources should be colocated in the same region to reduce latency and avoid cross-region egress charges. A Compute Engine instance in one region that frequently accesses a Cloud Storage bucket in another region creates a sustained cross-region flow that raises cost and slows responses, and Cloud Storage supports bucket relocation to move the bucket to the same region as the services that use it.
An infrastructure team is designing a highly available architecture on Google Cloud to withstand localized infrastructure failures. The architecture must satisfy the following architectural requirements:
Which combination of Google Cloud resource scopes correctly aligns with these requirements?
Last updated