Prepare and test your skills

Prepare and test your skills

Financial governance and managing cloud costs is the system of budgets, quotas, discounts, and reports that keeps Google Cloud spending visible, assigned to owners, and within limits. Budgets track planned versus actual spending and send alerts, quotas block excess resource creation, and billing reports connect every charge to a project, team, or business purpose. Together these controls let product owners, developers, administrators, and financial analysts make cost-effective decisions before costs grow.
Budgets, quotas, and spend caps are complementary guardrails where budgets warn about money and quotas and caps block consumption. A budget defines how much a team, project, service, or label should spend, while cost allocation assigns a portion of the total budget to each team so each team owns its spending. Spending and utilization data flows from running resources to the Billing dashboard and Cloud Monitoring, then the system sends an alert to administrators when a threshold is met, and finally administrators review the data and right-size resources.
A three-column comparison showing Budgets send alerts while services keep running, Quotas block extra resource creation, and Spend Cap budgets stop new usage at 100 percent until manually lifted.
Spend cap budgets are a stricter type of budget that automatically blocks new usage when gross, estimated costs exceed 100% of the budget amount in a single project. Alert emails are sent when spend exceeds 50% and 80% of the cap, while in-flight requests run to completion but no new on-demand or commitment-covered usage can start. Spend caps do not pause persistent resources such as compute and storage, normal service returns only after manual action in the billing console, and caps apply only to specific services such as Gemini API, Gemini Enterprise Agent Platform, Cloud Run, and Cloud Run functions in one project with a monthly period starting on the first day.
| Control | Primary purpose | What happens at the threshold or limit |
|---|---|---|
| Budgets | Track planned versus actual spending | Stakeholders receive an alert and then decide how to adjust |
| Quotas | Limit how much can be provisioned or consumed | Additional resource creation is blocked |
| Spend cap budgets | Pause eligible services when enforced | New usage stops until the cap is manually lifted |
Exam tip: Budgets notify stakeholders when spending thresholds are reached, while quotas prevent unexpected costs by blocking additional resource consumption.
Discounts and rightsizing lower the effective rate by matching resources to real demand and paying the lowest price for what is used. Committed Use Discounts (CUDs) give lower prices in exchange for committing to a minimum quantity or spend for one year or three years, with resource-based CUDs for Compute Engine in a specific region and spend-based CUDs for spend on a product at the billing account level. Sustained Use Discounts (SUDs) are automatic credits for continuous Compute Engine and Google Kubernetes Engine (GKE) use during a billing cycle with no commitment required.
Rightsizing fixes waste that discounts alone cannot fix because data flows from the monitored workload to Cloud Monitoring to Recommender and then to the administrator as a recommendation. Recommender targets idle Compute Engine instances that can be decommissioned, unneeded attached resources that still generate charges, and over-provisioned virtual machines (VMs) that need a cheaper machine type. Spot VMs use unused VM capacity at more than 80% lower cost than regular VMs, but Compute Engine can preemptively stop or delete them, so they fit only fault-tolerant and flexible workloads such as batch jobs.
| Option | How savings are earned | Choose when |
|---|---|---|
| Committed Use Discounts (CUDs) | Commit to minimum quantity or spend for one to three years | Usage is predictable and steady over a long term |
| Sustained Use Discounts (SUDs) | Run Compute Engine or GKE continuously past monthly thresholds | Usage is steady through the month but no long-term commitment is wanted |
| Spot VMs | Use unused capacity at more than 80% lower cost | Workload can tolerate preemption |
Hierarchy, labels, and billing reports create accountability by connecting every charge to an owner and a business purpose. Costs flow upward from individual resources to the project that contains them, then to folders and the Organization, while a Cloud Billing account tracks charges for all linked projects for payment. Ownership that controls Identity and Access Management (IAM) inheritance is separate from payment linkage that defines which billing account pays, and every project must be linked to a billing account to function.
Labels such as team=banking_appdev and tags are key-value pairs forwarded to billing so reports can break down costs by team, product, region, or environment. A strong metadata strategy keeps this accurate through validity that maps labels to business needs, automation with tools like Terraform to avoid manual errors, and simplicity that avoids confusing rules. Cloud Billing Reports can then be filtered by service, SKU, project, or label and grouped by project, service, SKU, location, label, or project hierarchy, with the Cost Table for invoice reconciliation, the Cost Breakdown waterfall for costs after savings, and export to BigQuery for custom dashboards and chargeback.
Charge period shows actual usage costs incurred during selected dates back to January 2017 and supports forecasts up to 12 months ahead, but it excludes invoice-level charges such as taxes. Billing period shows usage plus invoice-level charges for selected complete months back to January 2019, but it does not support forecasts. Forecasting follows an ordered flow from gathering historical spend to preparing data and handling outliers to accounting for seasonality and trends to structuring costs for stable time-series analysis.
Continue with this study guide: Financial governance and managing cloud costs
Operational excellence and reliability at scale means keeping applications working through failures, traffic spikes, and change without constant manual work. Reliability comes from spreading work across independent failure domains, automating detection and recovery, and protecting data with managed backup. Together these practices support scalable, secure, and continuously available services.
Redundancy across zones and regions is duplicating critical components so one failure does not stop the whole application. A failure domain is a level that can fail independently, from a single VM to a zone or a region, and a single point of failure is a component whose loss makes the whole stack unavailable. Location scope controls how far a failure can spread, and wider scope costs more in resources, cross-location traffic, and replication, so teams choose the lowest scope that meets business needs.
Zonal resources such as Compute Engine VMs must be managed for redundancy by the team, while regional resources such as regional managed instance groups and Cloud Run services spread automatically across zones. Managed multi-regional resources such as Cloud Storage, Spanner, Firestore, and Datastore are designed to keep working after the loss of a single region. Single-zone architectures target 99.9% availability, multi-zone architectures target 99.99%, and multi-region architectures target 99.999% for business-critical workloads.
High availability (HA) generally applies to regional deployments with synchronous replication between zones and active-active topology with near-zero recovery time objective (RTO) and recovery point objective (RPO). Disaster recovery (DR) generally applies to multi-region deployments and typically uses asynchronous replication with manual or semi-automated failover. RTO is the maximum acceptable downtime and RPO is the maximum acceptable data loss in time, and smaller RTO and RPO values cost more and add complexity.
Site Reliability Engineering (SRE) and DevOps are complementary ways to deliver software quickly without breaking reliability. SRE focuses on running reliable systems at scale through eliminating toil, monitoring everything, and blameless postmortems, while DevOps focuses on processes, culture, and tooling that break silos between development and operations through continuous integration and deployment automation. An application is reliable when it meets objectives for availability and resilience.
A service-level indicator (SLI) measures performance such as the ratio of successful responses to all responses, a service-level objective (SLO) sets a target for that measurement over a compliance period such as 99% availability over a calendar week, and a service-level agreement (SLA) is the uptime commitment Google makes for a service. The error budget is the allowed unreliability between 100% uptime and the SLO, such as 0.01% for a 99.99% SLO. While budget remains, teams can release features, and when the budget is depleted, teams pause releases and focus on stability.
Code moves through Google Cloud in a fixed order where each step depends on the previous step. First, code flows from the source repository to Cloud Build, which builds, tests, and deploys containers. Then the tested image flows from Cloud Build to Artifact Registry, which stores, manages, and secures it. Finally, the stored artifact flows from Artifact Registry to Cloud Deploy, which delivers it to target environments in sequence.
A left-to-right process flow from Source Repository to Cloud Build to Artifact Registry to Cloud Deploy, then sequentially to Dev, Staging, and Prod environments.
Exam tip: An SLI measures performance, an SLO sets a target for that measurement, and an SLA is the contractual commitment; the error budget is the gap between 100% and the SLO target.
Google Cloud Observability collects metrics, logs, traces, and errors in one place so teams detect problems, find causes, and fix them faster. Cloud Monitoring collects metrics, events, and metadata for dashboards and alerts, Cloud Logging stores and analyzes log data, Cloud Trace collects latency data for distributed tracing, and Error Reporting aggregates application errors. Telemetry flows from instrumented application code into Monitoring as metrics and into Logging as structured logs, and exemplars link a metric measurement to a representative trace.
Infrastructure as Code (IaC) defines infrastructure through code for consistent, repeatable, and testable deployments. Terraform provisions Google Cloud resources, Ansible, Chef, and Puppet configure resources after provisioning, and GitOps defines operations as code. Autoscaling then turns measured data into action where Compute Engine managed instance groups scale VM groups within minimum and maximum limits, GKE Cluster Autoscaler adds or removes nodes based on Pod demands, and Cloud Run scales to zero instances by default when there is no traffic.
Google Cloud Customer Care is the support portfolio that helps organizations adopt cloud faster and operate reliably through how-to advice, troubleshooting, and operational knowledge. Basic Support is included for all customers with documentation and billing support, while paid tiers add unlimited one-to-one technical support. Standard Support fits small to medium organizations with workloads under development, Enhanced Support fits medium to large companies with production workloads and 1-hour P1 response 24/7, and Premium Support fits enterprises with 15-minute P1 response 24/7 plus a Technical Account Manager (TAM).
A TAM is a strategic advisor included with Premium Support that focuses on onboarding, cost management, adoption roadmaps, and preparation for critical events. Operational Health Reviews are regular TAM-led touchpoints that cover support and spend trends, incidents and outages, and high-priority projects. Mission Critical Services adds proactive monitoring and Priority 0 (P0) filing with 5-minute response for environments where disruption would cause significant impact.
The life of a support case moves from creation and triage through communication to resolution and learning. Creation follows a fixed flow:
A new case starts as New when unassigned, moves to Assigned when a specialist takes ownership, continues as In progress during specialist or engineering work, and may shift to Waiting on customer response when more information is needed. P1 means critical impact with unusable production service, P2 means high impact with severely impaired service, P3 means medium impact, and P4 means low impact. A closed case can be reopened only within 15 days.
Continue with this study guide: Operational excellence and reliability at scale
Sustainability with Google Cloud covers measuring cloud emissions and then lowering them through cleaner regions and more efficient workloads. Measurement turns energy use into trackable emissions data, while reduction removes waste from compute and storage. Together they link lower carbon with lower cost.
The Carbon Footprint dashboard shows estimated greenhouse gas emissions from covered Google Cloud services for a selected billing account. Every resource created in Google Cloud has an associated footprint, and the dashboard aggregates that impact at the billing account level. Data is computed automatically with no setup, but previous-month data can take up to 21 days to appear.
Access lives in the Tools section of the Google Cloud console and is controlled by IAM on the billing account. A billing account administrator grants Carbon Footprint Viewer, Billing Admin, or Billing Viewer, and viewing carbon data requires carbon permissions but does not require access to billing spending data. Summary cards show an annual footprint for the past 12 complete months and a past-month figure compared with the previous month, with charts that break emissions down by month, by region, and by project and product.
Market-based and location-based views are the two Scope 2 accounting methods in the Carbon Footprint dashboard. Both follow the Greenhouse Gas Protocol, and Scope 1 and Scope 3 emissions are identical on both tabs. Only the Scope 2 calculation differs between them, with emissions shown in metric tons of CO2 equivalent (tCO2e) in the dashboard and kilograms of CO2 equivalent (kgCO2e) in the export.
| Tab | Scope 2 method | What it shows |
|---|---|---|
| Market-based emissions | Scope 2 after carbon-free energy purchases | Emissions by scope 1, scope 2 market-based, and scope 3 |
| Location-based emissions | Scope 2 from the local grid | Emissions by scope 1, scope 2 location-based, and scope 3, plus breakdowns by project and by product |
Exporting Carbon Footprint data to BigQuery enables custom analysis beyond the built-in dashboard. The BigQuery Data Transfer Service creates a monthly partitioned table named carbon_footprint in a chosen dataset. Exports run on the 15th day of the following month, backfill can reach back to January 2021, and BigQuery storage and query use is billed. Once data flows from Carbon Footprint to BigQuery, teams can build views in Looker Studio and join carbon exports with billing data to estimate emissions by instance, tag, or label.
Exam tip: Market-based and location-based tabs differ only in how Scope 2 emissions are calculated; Scope 1 and Scope 3 emissions are identical across both tabs.
Shared responsibility for sustainable architecture splits work between Google and the customer at the boundary between platform and workload. Google owns sustainability of Google Cloud, including efficient data centers with high power usage effectiveness (PUE), renewable energy, climate-conscious cooling, and hardware optimization. The customer owns sustainability in the cloud, including right-sizing resources, using serverless services that scale to zero, and managing data lifecycles.
Low-carbon region choice places workloads where carbon-free energy percentage is higher or grid carbon intensity is lower. Teams compare published carbon-free energy scores and carbon-intensity data, then enforce the choice with the resource locations constraint in Organization Policy Service, which blocks new resources in high-carbon regions. Migrating on-premises workloads to Google Cloud can cut energy use and associated emissions by 1.4 to 2 times because large cloud data centers use modern cooling and power distribution.
Efficiency actions remove energy waste from compute and storage and depend on keeping compute near data so data flows from storage to compute without crossing regions. Intelligent workload management uses Active Assist machine type recommendations, aggressive autoscaling, and serverless scaling to zero so no energy is spent waiting. Storage optimization uses lifecycle rules to transition infrequently accessed data to Nearline, rarely accessed data to Coldline, and long-term retention data to Archive, plus compressed columnar formats like Parquet, deletion of old snapshots, and limits on cross-region replication to mandatory DR or HA needs.
Continue with this study guide: Sustainability with Google Cloud
Budgets track planned versus actual spending and send alerts so stakeholders can decide how to adjust, quotas limit how much can be provisioned or consumed by blocking additional resource creation, and spend cap budgets pause eligible services by stopping new usage until the cap is manually lifted. Spend cap budgets block new usage when gross estimated costs exceed 100 percent of the budget amount in a single project, with alert emails at 50 percent and 80 percent, while in-flight requests run to completion and persistent resources such as compute and storage are not paused.
Committed Use Discounts (CUDs) give lower prices in exchange for committing to a minimum quantity or spend for one year or three years and fit predictable and steady long-term usage. Sustained Use Discounts (SUDs) are automatic credits for continuous Compute Engine and Google Kubernetes Engine (GKE) use during a billing cycle with no commitment required. Spot VMs use unused capacity at more than 80 percent lower cost than regular VMs but Compute Engine can preemptively stop or delete them, so they fit only fault-tolerant and flexible workloads such as batch jobs.
A service-level indicator (SLI) measures performance such as the ratio of successful responses to all responses, a service-level objective (SLO) sets a target for that measurement over a compliance period such as 99 percent availability over a calendar week, and a service-level agreement (SLA) is the uptime commitment Google makes for a service. The error budget is the allowed unreliability between 100 percent uptime and the SLO, such as 0.01 percent for a 99.99 percent SLO, and while budget remains teams can release features while a depleted budget means teams pause releases and focus on stability.
Market-based emissions and location-based emissions differ only in how Scope 2 emissions are calculated, while Scope 1 and Scope 3 emissions are identical across both tabs. The market-based view shows Scope 2 after carbon-free energy purchases and the location-based view shows Scope 2 from the local grid, with emissions shown in metric tons of CO2 equivalent in the dashboard and kilograms of CO2 equivalent in the export.
Last updated