Cost Optimisation

The Cost Optimisation pillar of the GCP Architecture Framework focuses on maximising the value you get from every pound or dollar spent on Google Cloud. Cost optimisation is not about spending less — it is about spending wisely. The goal is to deliver the same (or better) outcomes at the lowest possible cost, without sacrificing reliability, security, or performance.

---

Design Principles

Principle	Description
Measure and attribute costs	You cannot optimise what you do not measure — implement cost visibility first
Right-size resources	Match resource allocation to actual workload requirements
Use committed discounts	Lock in savings for predictable workloads with committed use discounts
Leverage serverless	Pay only for what you use with services like Cloud Run and Cloud Functions
Eliminate waste	Identify and remove idle, orphaned, and over-provisioned resources
Architect for cost	Consider cost implications in architectural decisions from the start

---

Cost Visibility

Cloud Billing

Feature	Description
Billing accounts	Link projects to billing accounts for consolidated invoicing
Billing export to BigQuery	Export detailed billing data for custom analysis and dashboards
Budget alerts	Set spending thresholds with notifications at configurable percentages
Cost breakdown	View costs by project, service, SKU, and label

Billing Export to BigQuery

# Enable billing export to BigQuery (configured in Console)
# Then query your costs:

SELECT
  project.name AS project,
  service.description AS service,
  SUM(cost) AS total_cost,
  SUM(usage.amount) AS total_usage
FROM `my-billing-project.billing_export.gcp_billing_export_v1_XXXXXX`
WHERE DATE(usage_start_time) >= DATE_SUB(CURRENT_DATE(), INTERVAL 30 DAY)
GROUP BY project, service
ORDER BY total_cost DESC
LIMIT 20

Labels for Cost Attribution

# Apply labels to resources for cost tracking
gcloud compute instances update my-vm \
  --update-labels=team=platform,environment=production,cost-centre=engineering

# Label a GKE cluster
gcloud container clusters update production-cluster \
  --region=europe-west2 \
  --update-labels=team=backend,product=orders

Budget Alerts

# Create a budget with alert thresholds
gcloud billing budgets create \
  --billing-account=0X0X0X-0X0X0X-0X0X0X \
  --display-name="Production Budget" \
  --budget-amount=5000 \
  --threshold-rule=percent=0.5 \
  --threshold-rule=percent=0.8 \
  --threshold-rule=percent=1.0

---

Right-Sizing

Right-sizing means matching your resource allocation to actual workload requirements — neither over-provisioned (wasting money) nor under-provisioned (degrading performance).

GCP Recommender

# View VM right-sizing recommendations
gcloud recommender recommendations list \
  --project=my-project \
  --location=europe-west2-a \
  --recommender=google.compute.instance.MachineTypeRecommender

# View idle VM recommendations
gcloud recommender recommendations list \
  --project=my-project \
  --location=europe-west2-a \
  --recommender=google.compute.instance.IdleResourceRecommender

Custom Machine Types

GCP allows you to create VMs with custom vCPU and memory configurations, avoiding the waste of standard machine types:

# Create a VM with exactly the resources you need
gcloud compute instances create custom-vm \
  --custom-cpu=6 \
  --custom-memory=20GB \
  --zone=europe-west2-a

---

Committed Use Discounts (CUDs)

CUDs provide significant savings for predictable workloads in exchange for a 1-year or 3-year commitment:

Commitment	1-Year Discount	3-Year Discount
Compute	Up to 37%	Up to 55%
Memory-optimised	Up to 37%	Up to 55%
Cloud SQL	Up to 25%	Up to 52%

# Purchase a committed use discount
gcloud compute commitments create my-commitment \
  --region=europe-west2 \
  --plan=36-month \
  --resources=vcpu=100,memory=400GB

CUD vs Sustained Use Discounts