Cloud cost & FinOps glossary

Plain-language definitions of the cloud cost and FinOps terms that show up on your bill and in optimization conversations — from egress and NAT gateways to reserved instances, amortized cost, and chargeback.

Amortized cost

Amortized cost spreads upfront commitment payments (like reserved instances or Savings Plans) evenly across the hours they cover, showing the true effective cost of usage over time.

Autoscaling

Autoscaling automatically adds or removes capacity in response to demand, so you pay for resources when they are needed and release them when they are not.

Blended vs unblended cost

Unblended cost is the actual rate charged for each usage record, while blended cost averages rates across an organization's linked accounts — the two can differ when reservations are shared.

Block storage

Block storage provides raw disk volumes (like AWS EBS, Azure Managed Disks, or GCP Persistent Disk) attached to instances, priced by provisioned gigabytes and performance tier.

Chargeback

Chargeback bills cloud costs directly back to the teams or business units that incurred them, making cloud spend a line item in their own budgets.

Cloud waste

Cloud waste is spending on cloud resources that deliver no value — idle instances, oversized machines, orphaned storage, and forgotten environments.

Committed use discount

A committed use discount is Google Cloud's equivalent of reserved pricing — a 1- or 3-year commitment to a level of resource usage in return for a lower rate.

Cost allocation

Cost allocation is the practice of attributing cloud spend to the teams, products, or features that generate it, usually through resource tagging and account structure.

Cost and Usage Report

The Cost and Usage Report (CUR) is AWS's most detailed billing dataset, providing line-item, hourly usage and cost data for every resource in an account.

Data egress

Data egress is data leaving a cloud provider's network to the internet or another region, and it is one of the most commonly underestimated line items on a cloud bill.

FinOps

FinOps is the practice of bringing financial accountability to variable cloud spending, so engineering, finance, and product teams make cost-aware decisions together.

Idle & orphaned resources

Idle and orphaned resources are provisioned but unused assets — stopped instances still holding disks, unattached volumes, unused IP addresses, and old snapshots — that keep billing for no benefit.

Load balancer

A load balancer distributes incoming traffic across multiple instances, and cloud-managed load balancers bill an hourly base charge plus a capacity/throughput fee.

NAT gateway

A NAT gateway lets resources in a private subnet reach the internet for outbound connections, and it bills both an hourly rate and a per-GB data-processing fee.

Object storage

Object storage keeps data as objects in flat buckets (like Amazon S3, Azure Blob, or Google Cloud Storage), priced by gigabyte stored per month plus request and retrieval fees.

On-demand pricing

On-demand pricing charges for compute by the second or hour with no commitment, at the highest per-unit rate — the default and most flexible way to pay for cloud resources.

Reserved instance

A reserved instance is a 1- or 3-year commitment to a specific instance configuration in exchange for a large discount (up to ~72%) versus on-demand pricing.

Resource tagging

Tagging attaches key-value metadata (like team, environment, or cost-center) to cloud resources so spend can be grouped, filtered, and allocated.

Rightsizing

Rightsizing is matching an instance or resource to the smallest size that still meets its performance needs, eliminating the cost of overprovisioned capacity.

Savings Plan

A Savings Plan is an AWS commitment to a steady hourly dollar amount of compute usage for 1 or 3 years, giving reserved-level discounts with more flexibility than classic reserved instances.

Showback

Showback reports each team or product's cloud costs to make spending visible, without actually billing those costs back to their budgets.

Spot instance

A spot instance is spare cloud capacity sold at a steep discount (often 70–90% off on-demand) that the provider can reclaim with little notice.

vCPU

A vCPU is a virtual CPU — a share of a physical processor core allocated to a cloud instance, and the primary unit cloud providers use to size and price compute.