Plain-English definitions of the terms that show up on every cloud bill and in every FinOps conversation. Written buyer-side, the way we explain them to the teams we work with.
Automated alerting that flags cloud spend deviating from its expected pattern, ideally the day it happens rather than at month end. A core part of the Lock step in our method, alongside budgets and guardrails.
Capacity that follows demand up and back down rather than sitting at peak all day. Done well it cuts cost without raising outage risk; see autoscaling done right.
The way committed-purchase costs are spread across the term they cover, so a one-time reservation payment shows up smoothly month by month rather than as a single spike. Essential for reading a bill that includes commitments.
An agreement to use a baseline amount of cloud capacity or spend for one to three years in exchange for a lower rate. Reserved Instances, Savings Plans and Committed Use Discounts are all commitments. Buy them on a rightsized baseline, never before.
Google Cloud's commitment instrument, discounting compute and some other services in return for a one or three year usage commitment.
Assigning every dollar of cloud spend to an owner, team, environment and service, usually through tagging. You cannot optimize what you cannot attribute. See tackling untagged and unowned resources.
The tendency for large, central datasets to pull applications and other data toward them, raising both architecture complexity and cost over time. See the hidden cost of data gravity.
The charge for moving data within or out of a cloud provider. The internal kind, between regions and zones, is often larger than the internet-facing kind. See reducing inter-region data transfer costs.
The cost of data leaving a cloud provider or crossing a billed boundary. One of the most common surprises in a bill review because it is designed in, not chosen. See data egress charges explained.
AWS's negotiated agreement that trades a multi-year committed spend for a percentage discount across the whole bill. Similar private-pricing agreements exist on other clouds.
The operating practice of bringing financial accountability to the variable spend of the cloud, organized around the phases Inform, Optimize and Operate. Our See, Cut, Lock, Run method maps onto it.
The FinOps Open Cost and Usage Specification, an open standard that normalizes billing data from different providers into one schema so AWS, Azure, GCP and OCI can be read in a single currency of cost.
A resource that is running and billing but doing little or no useful work, such as a weekend development instance or a load balancer with no healthy targets. See finding idle cloud resources.
A pricing class for stored objects that trades retrieval speed and cost for storage price: hot, cool, cold and archive. Matching data to the right tier is the largest single storage saving. See object storage tiers compared.
Running more capacity than a workload uses, usually because engineers size for peak and forget to scale back. The structural cause of most rightsizing opportunity; see why over-provisioning happens.
A commitment to a specific instance type in a region for one or three years in exchange for a lower rate. The oldest commitment instrument, now often paired with or replaced by Savings Plans.
Matching a resource to the capacity its workload actually uses, rather than the capacity it was provisioned with. The fastest-converting optimization. See the complete guide to cloud rightsizing and waste elimination.
A flexible commitment to a fixed hourly spend on compute in exchange for a lower rate, available on AWS and Azure. More flexible than a Reserved Instance because it is not tied to a single instance type.
A point-in-time copy of a disk or volume, used for backup. Snapshots accumulate forever unless something prunes them. See snapshot and backup cost optimization.
An automated rule that moves objects to cheaper tiers and deletes them on a schedule, with no human in the loop. Build once, save forever. See building a storage lifecycle policy.
Labeling cloud resources with metadata, such as team, environment and cost center, so spend can be allocated. Tags applied by hand decay, so a durable strategy enforces them automatically.
The cost of the cloud per unit of business value, such as dollars per customer or per transaction. A falling unit cost while absolute spend grows is the real sign that optimization is working.
Resources fully detached from anything that needs them: orphaned disks, unattached IPs, snapshots of deleted instances. Pure waste, and a prime target for elimination. See finding zombie infrastructure.
Our cloud cost audit reads your spend across AWS, Azure, GCP and OCI and ranks every opportunity by dollars. On the performance model, you pay only from realized savings.
Book a cloud cost audit →