PerfectScale Pricing for the Free, Advanced and Expert Plans Compared
By the CostAnalyst team
projected this month if unattended
Spend by team
Budget forecast
PerfectScale by DoiT is free up to 300 vCPUs a month and quote-only above that. Its pricing page shows two paid plans, Advanced and Expert, both billed per vCPU, with no dollar figure on either. The only numbers in public are third-party: a directory lists Advanced at $2 and Expert at $4 per monthly vCPU, with a 200 vCPU minimum on Expert. A PerfectScale quote is therefore negotiated from a rate you cannot confirm in writing, and the sensible way to budget is to model the range and ask for the written rate before the pilot ends.
That gap matters because PerfectScale sits in the Kubernetes rightsizing category, where the vendor and the buyer rarely meet on a price sheet. Everything below was read from perfectscale.io and the DoiT help center on 31 October 2026. Where a number comes from somewhere else, the page says so.
How much does PerfectScale cost?
PerfectScale has four plan names in its own documentation: Trial, Community, Advanced and Expert, plus a custom Enterprise conversation. The public pricing page shows three columns.
| Plan | What perfectscale.io shows | Third-party listing |
|---|---|---|
| Free (Community) | Free up to 300 vCPU a month, with a sign-up button | Free for life, up to 300 vCPUs |
| Advanced | "per vCPU", button reads Get a quote, no figure | $2 per monthly vCPU |
| Expert | Marked Most Popular, "per vCPU", Get a quote, no figure | $4 per monthly vCPU, minimum 200 vCPUs |
| Enterprise | Custom | On request |
The $2 and $4 figures come from a software directory, not from PerfectScale, and the directory itself tells readers to confirm before buying. Treat them as a planning range. PerfectScale has had an AWS Marketplace listing since July 2023, but we could not re-read a posted rate from it on the day of writing, so the figures here are a range to confirm, not a first-party card.
How does PerfectScale bill a vCPU?
PerfectScale bills on average monthly vCPU, not on peak and not on node count. The DoiT help center gives the formula: total vCPU-hours used during the month divided by the total hours in that month, which is the average concurrent vCPU across the whole month. A cluster that runs 400 vCPUs for half the month and 100 for the other half bills as 250.
That is friendlier to autoscaling clusters than a peak-based meter, and it means the tool's own savings lower its own bill. A team that cuts requests by 30 percent also cuts the vCPU count it pays on the following month.
What does PerfectScale cost at 500 and 2,000 vCPUs?
Using the directory's rates and assuming the whole fleet is billed once you pass the free allowance (PerfectScale does not say whether the first 300 are netted off):
| Average monthly vCPUs | Advanced at $2 | Expert at $4 |
|---|---|---|
| 250 | $0 (inside the free allowance) | Not applicable, 200 minimum |
| 500 | $12,000 a year | $24,000 a year |
| 2,000 | $48,000 a year | $96,000 a year |
If the first 300 vCPUs are free inside a paid plan, subtract $7,200 a year from the Advanced rows and $14,400 from the Expert rows. That is the first question to put in writing.
Does PerfectScale for Commitments cost extra?
Yes, and it is priced differently. PerfectScale for Commitments, the Savings Plans product DoiT launched alongside the Kubernetes tool, charges a percentage of savings on customer-owned Compute and Database Savings Plans. The DoiT documentation defines the fee as on-demand cost minus Savings Plan cost, times your contracted percentage, calculated daily. The percentage is set in the contract and is not published, and the page states no minimum fee. Flexsave Savings Plans are excluded.
For context, a savings-share model of this kind is how Zesty prices Commitment Manager and how Archera prices insured commitments. A fee on savings is cheap when commitments are small and expensive once they are large, so ask for a cap.
Is the PerfectScale free plan enough?
For a single cluster or a few small ones under 300 vCPUs, yes, as a way to see recommendations on your own workloads. It is not a way to run production automation. The vendor's pricing page states the limit but not which features the free tier includes, and the plan comparison checkmarks do not render in the page text, so confirm that automatic rightsizing is on the plan you intend to run. Beyond 300 vCPUs you are in a sales conversation either way.
How does PerfectScale pricing compare to its alternatives?
| Tool | Published pricing | Where to read it |
|---|---|---|
| PerfectScale by DoiT | Free to 300 vCPUs, paid per vCPU with no public figure | This page |
| Zesty Platform | $50,000 a year on AWS Marketplace | Zesty pricing |
| ScaleOps | $50,000 platform fee plus $9.00 per vCPU on AWS Marketplace | ScaleOps pricing |
| StormForge Optimize Live | $0.0041 per requested vCPU-hour, or $32,400 a year per 1,000 vCPUs | StormForge pricing |
| CAST AI | From $1,000 a month plus $5 per CPU on AWS Marketplace | CAST AI pricing |
On the numbers that exist, PerfectScale is the cheapest of the group at 500 vCPUs on the Advanced rate, and it is also the only one where the price cannot be confirmed from the vendor. The full side-by-side of features is in our Kubernetes rightsizing tools comparison.
What should you ask PerfectScale before signing?
- What is the written per-vCPU rate on Advanced and Expert, and is it the same at 500 and 2,000 vCPUs?
- Are the first 300 vCPUs free inside a paid plan, or does the whole fleet bill once you pass the limit?
- Which features are on Advanced and which are Expert only, in particular automatic rightsizing and the accelerators?
- Is the term monthly, annual or multi-year, and what are the renewal and cancellation terms? A contract clause review tool is worth running over the order form if renewal language is unclear.
- If you buy Commitments, what is the savings percentage and is there a cap?
- Can the purchase draw down a cloud marketplace commitment through DoiT?
The short version
PerfectScale is free to 300 vCPUs a month and billed on average monthly vCPU above that, with Advanced and Expert plans that carry no public price. A directory lists $2 and $4 per vCPU, which gives roughly $12,000 to $24,000 a year at 500 vCPUs, but get the rate in writing before relying on it. PerfectScale for Commitments takes a contracted share of Savings Plan savings. If what you need first is to see where cloud and SaaS spend is going before buying a rightsizing tool, CostAnalyst's plans start at $59 a month billed yearly, and the one-time analysis shows the waste in a single billing file.
See where your cloud and SaaS money is leaking
Upload your cloud billing export and see your savings in dollars. Transparent pricing, no card to start.