FAQ¶
Common questions about cloud cost optimisation, the FinOps AI calculators, and the safe-remediation approach.
About the calculators¶
Are the numbers from the calculators accurate?¶
The math is sound, but the inputs are illustrative. Specifically:
- EC2 hourly rates come live from instances.vantage.sh (us-east-1, Linux, on-demand) — they reflect AWS's current public price book.
- Service prices (NAT Gateway $/hour and $/GB, EBS gp2/gp3 $/GB-month, EKS control plane $/hour) are hardcoded illustrative pegs sourced from AWS's public pricing pages at the time of writing.
Your real AWS invoice depends on region, negotiated commitment (EDP / PPA), storage class, and data-transfer destination. Use the calculators to get a directionally correct estimate, and book a free audit when you want a verified number against your real bill.
Do the calculators store my inputs?¶
No. The math runs entirely in your browser. We don't send your inputs to a server, and we don't log them. The audit form is the only place where you'd choose to share data with us — and only when you click "Request my free audit".
Why a calculator per topic instead of one big one?¶
Because the math is different for each lever. The Cloud Waste Radar is the "give me the aggregate number" calculator; the per-topic ones (NAT, gp2 → gp3, EKS) go deep on one cost lever each.
About AWS cost optimisation in general¶
What's the single highest-ROI fix?¶
Almost always adding S3 + DynamoDB Gateway endpoints to your VPCs. They are free, they take effect immediately, and they typically remove 50-60% of the NAT Gateway data-processing slice on day one.
Is gp2 → gp3 really always worth it?¶
For ~95% of workloads, yes. The exceptions are:
- Very small (sub-100-GB) volumes where the savings are pennies/month.
- Workloads that sustain >3,000 IOPS without bursting — they need provisioned IOPS on
gp3, which can match or exceed gp2's cost depending on the volume.
For most teams, the answer is "migrate everything except the database that's already on io2".
What's the right number of EKS clusters?¶
For most startups: 3 (dev / staging / prod), maybe 4 if you have a regulated workload that needs isolation. More than that is usually a smell. "One cluster per team" or "one cluster per service" is the canonical anti-pattern.
How quickly can a team realise the savings?¶
In our audits, the typical sequence is:
- Week 1: VPC endpoints + CloudWatch retention + gp2 → gp3 of dev/staging — usually 30–50% of the projected savings, delivered in five small PRs.
- Month 2: EKS consolidation + RDS right-sizing — another 20–30%.
- Month 3: Compute Savings Plans + AZ-pinning audit — the remaining 20-30%.
A motivated platform team can do the whole quarter plan in eight weeks. The bottleneck is usually approvals, not engineering.
About FinOps AI (the product)¶
What does the FinOps AI product actually do?¶
It's a safe-remediation operator for AWS and GCP. The calculators on this site estimate what's possible; the product executes the savings actions against your cloud account — with typed approval flows, blast-radius guardrails, pre- / post-action health checks, signed audit trail, and automatic rollback if a regression is detected.
In one sentence: Vantage and Finout show you what to do; FinOps AI does it.
→ See how it works on the main site
How is it different from Vantage / Finout / CloudHealth?¶
The recommendation engines are similar in coverage — idle EC2, unattached EBS, snapshots, RDS right-sizing, NAT-endpoint candidates, gp2 → gp3, EKS waste. The difference is the next step:
- Vantage / Finout hand the recommendation back to your team to act on.
- FinOps AI executes the approved action with a kill switch.
→ See the side-by-side comparison
What level of access does the FinOps AI audit need?¶
Read-only. The free audit attaches an IAM role with a curated read-only policy (Cost Explorer, EC2 describe-, RDS describe-, EKS describe-, S3 list-, CloudWatch get-*). No write permissions, no production data access. The audit deliverable is a written report, not a code change.
If you decide to proceed with the product, you control which write permissions you grant for the safe-remediation actions. Each action is approved with a typed change-request; nothing runs without your team's go.
Booking an audit¶
What happens after I submit the audit form?¶
We reply within one business day with two suggested time slots (30-minute call) and a short list of what we'll need (the read-only IAM role + access to your AWS Organisation's monthly cost CSV). No NDA required for the initial call; we sign mutual NDAs before any data exchange if you want one.
What if I don't have an AWS Organisation?¶
A single AWS account is fine — most of our audits are single-account. Multi-account work is more complex but the deliverable shape is the same.
Is there an obligation to buy the product?¶
No. The audit deliverable is yours — a written report with the prioritised savings list. Most teams implement the top 3–5 items themselves; we're available to help when you want to.
Still have a question?¶
Email us at hello@getfinops.cloud or book a free 30-minute audit and we'll start with your specific situation.