CDL
Deep Dive
Billing, budgets and alerts, discounts (SUD, CUD, Spot VMs), Google Cloud Pricing Calculator, and FinOps on GCP.
GCP organizes billing in a hierarchy: the Billing Account pays for one or more projects. An organization can have multiple billing accounts.
Cloud Billing
Console for viewing invoices, exporting cost data to BigQuery, managing payment methods, and viewing payment history.
Cost Table
Detailed view of costs by project, service, or SKU over a period. Exportable to CSV.
Budget Alerts
Define budgets and receive alerts when actual or forecasted spend exceeds certain thresholds.
Billing Export → BigQuery
Exports detailed billing data to BigQuery for advanced analysis with SQL, Looker Studio, etc.
Sustained Use Discounts (SUD)
Automatic discounts for Compute Engine VMs that run for more than 25% of the month; the discount grows the longer the VM runs. No upfront commitment — GCP applies them automatically on the invoice.
Committed Use Discounts (CUD)
A commitment to purchase resources for 1 or 3 years in exchange for significant discounts. Two types: resource-based (vCPUs/memory) or spend-based (committed $).
Spot VMs
Low-cost VMs that GCP can interrupt with 30 seconds notice when it needs the capacity. They replace the older preemptible VMs (which lasted at most 24h): no time limit, but still interruptible.
| Tool | What it does |
|---|---|
| Google Cloud Pricing Calculator | Estimates the monthly cost before provisioning resources. You configure the services you plan to use and get a breakdown of estimated costs. |
| Budget Alerts | Define budgets and receive alerts (email, Pub/Sub) when actual or forecasted spend exceeds 50%, 90%, or 100% of the budget. |
| Recommender | Automatic suggestions to optimize resources: remove idle VMs, downsize underutilized VMs, remove unused reserved IPs. |
| Cost Breakdown | Cost breakdown by project, label, region, or service. Helps identify what generates the most spend. |
| Labels and Tags | Tag resources (VMs, buckets, clusters) with metadata (team, environment, project) to attribute costs in the invoice exported to BigQuery. |
| Quotas and limits | Set maximum resource usage limits to avoid unexpected costs from runaway scripts or misconfigurations. |
Rightsizing
Adjust VM size according to actual usage. Compute Engine Recommender identifies over-provisioned VMs. Downsize from n2-standard-8 to n2-standard-4 if average CPU is <20%.
Remove orphaned resources
Unattached static IPs, disconnected persistent disks, old snapshots, load balancers with no traffic. They generate cost even when not actively used.
Use managed resources
Cloud SQL, Cloud Spanner, and BigQuery scale and shut down based on usage. Always-on VMs have a fixed cost. Prefer PaaS over IaaS where possible.
Multi-tier storage architecture
Active data in Standard, monthly data in Nearline, yearly data in Coldline, archived files in Archive. Up to 95% reduction in storage cost for cold data.
Google matches its annual electricity use with purchases of renewable energy, and its data centers are among the most efficient (PUE close to 1.1). By 2030, Google aims to operate on 24/7 carbon-free energy across all its datacenters.
Using Google Cloud instead of an on-premises datacenter can reduce the carbon footprint of computing workloads by up to 90%.
100% renewable energy
Google has matched 100% of the energy it consumes with renewable energy purchases (Power Purchase Agreements) since 2017. It is the largest corporate purchaser of renewable energy in the world.
Carbon Footprint Tool
A free tool in the GCP console that shows the carbon emissions of your workloads (scopes 1, 2, and 3). Lets you identify which projects and services generate the most emissions.
Net zero goals 2030
Goal: operate on 24/7 carbon-free energy by 2030. Not just buying offsets — but ensuring that when you use GCP at 3am, the energy it uses is genuinely renewable at that moment.
Efficient PUE
Google's average PUE (Power Usage Effectiveness): 1.10. The industry average is 1.58. This means Google uses only 10% extra energy for cooling, while the industry uses 58%.
Circular economy in hardware
Google maximizes the useful life of its datacenter hardware and reuses or recycles components at end of life. Custom chip design (Titan, TPUs) for greater energy efficiency.
For the CDL exam: sustainability as a competitive advantage
Google Cloud positions sustainability as a business benefit, not just an environmental one. Companies with ESG (Environmental, Social, Governance) goals can use GCP to meet their emissions reduction targets and report them with the Carbon Footprint Tool. It is a differentiator against AWS and Azure.
Did you get this topic?
Test what you just learned
A company runs data processing jobs on Compute Engine for 8 hours every night (from 00:00 to 08:00). If the job fails partway through, it can restart from a checkpoint. Which type of VM reduces the cost the most for this use case?
Put what you learned to the test with original CDL exam questions.
Start CDL mock examUnlimited mock exams with the Pro plan
Log in to track your progress.