Skip to content

SolutionsFinOps

Know what your cloud costs, and why.

We make AWS spend traceable to the team that runs it, flag anomalies as they appear, and fix waste with your engineers.

Talk to a FinOps engineer48-hour Well-Architected Review

Definition

What is FinOps?

The practice of making cloud spend a shared job for engineering, finance and the business, so each decision weighs cost against value. The FinOps Foundation frames it in three phases: Inform, Optimise and Operate. On AWS it starts with clean cost data, tagged by owner.

Credentials

AWS Services Partner
Advanced Tier
AWS service deliveries
8
AWS certifications
25+
AWS certified
Every engineer

Where does cloud spend leak?

Most waste hides in a few familiar places. Each one has an AWS tool that finds it.

  1. Oversized instances

    Sized for a peak that never came.

    Found by AWS Compute Optimizer

  2. Unattached volumes

    EBS volumes and snapshots left behind.

    Found by Cost Optimization Hub

  3. Since Feb 2024

    Public IPv4 addresses

    Charged by the hour, in use or idle.

    Found by Amazon VPC IPAM

  4. NAT and cross-zone traffic

    Data that could stay in one zone.

    Found by CUR 2.0 · usage type

  5. Stale commitments

    Savings Plans sized for last year.

    Found by AWS Cost Explorer

  6. Spend nobody owns

    Untagged resources, no team to ask.

    Found by Tag policies

What does your team see each month?

Cost overview · AWS Organization · last 30 daysIllustrative data

Monthly spend

€48.2k

−12% vs last month

Spend with an owner tag

94%

up from 71%

Savings Plans coverage

68%

compute

Spend by owner tag · EUR / month

  • payments€19.4k
  • data-platform€11.2k
  • web€6.1k
  • shared€4.6k
  • untagged€2.9k
Illustrative numbers, not a client result.

How does waste actually get fixed?

Alerts go to the owner, and recommendations become code changes your engineers review.

Detect

Anomalies reach the owner, not a shared inbox.

AWS Cost Anomaly Detection watches each team's spend and names the service, account and usage type.

AWS Cost Anomaly DetectionMonitor · data-platform

Severity: medium

Unusual spend on NAT gateway data processing

Resourceeu-west-1 · NatGateway-Bytes

  1. Day 1Anomaly detected against the monitor's baseline
  2. Day 1Alert routed to the data-platform team
  3. Day 2Cause: nightly export reads Amazon S3 through NAT
  4. Day 3S3 gateway endpoint added; spend back to baseline

ResolvedFixed in Terraform; the monitor stays on.

Optimise

Recommendations become pull requests.

AWS Cost Optimization Hub gathers rightsizing, idle and Savings Plans advice. The ones that hold up become reviewed changes.

services/reporting/compute.tfTerraform+2−2
resource "aws_instance" "reporting" {Removed:   ami           = data.aws_ami.al2023_x86.idAdded:   ami           = data.aws_ami.al2023_arm64.idarm64Removed:   instance_type = "m5.2xlarge"Added:   instance_type = "m7g.large"Graviton  tags = {    owner = "data-platform"  }}
Policy checkCompute Optimizer findingFAIL · Overprovisioned→PASS · Optimized

How does the practice run?

  1. Step 01

    Inform.

    CUR 2.0 via AWS Data Exports, tags, cost by owner.

  2. Step 02

    Optimise.

    Rightsizing, clean-up and commitments as reviewed changes.

  3. Step 03

    Operate.

    Budgets, anomaly alerts, a monthly review with finance.

FAQ

Common questions.

With the data. We set up Cost and Usage Reports (CUR 2.0) through AWS Data Exports and a tagging policy, so every cost has an owner. Idle resources, unattached volumes and unused public IPv4 addresses are often the first fixes.

We size commitments from your usage history and recommend them; you approve every purchase. We revisit coverage as your baseline changes, so commitments follow the workload.

A review looks at cost once, alongside five other pillars. FinOps is the ongoing practice. Many teams start with the 48-hour AWS Well-Architected Review and continue here.

They should not. Guardrails live in code and alerts go to the owning team. Engineers see what their services cost without asking for approval on every change.

Next step

Find out where your spend goes.

Tell us how your AWS spend is organised. We show you where to look first.

Talk to a FinOps engineer