ReferenceBilling
Billing
How Kubeadapt charges, what counts toward your bill, and what happens between invoices. Per-vCPU pricing, the first 100 vCPU always free.
How pricing works
Kubeadapt is priced per vCPU, per month, against the workload requests in your cluster. Standard pricing is tiered against billable vCPU and drops as you scale. No per-feature pricing, no per-cluster fees, no per-seat charges. Every plan unlocks every feature.
Standard
From $1.99
/ vCPU / month · tiered
Startups
No tiers$0.60
/ vCPU / month · flat
Always free
0
first 100 vCPU
Standard tier schedule
- Billable 0–500$1.99
- Billable 500–2,000$1.49
- Billable 2,000–5,000$0.99
- Billable 5,000+$0.60
Each tier prices only the slice of billable vCPU inside it — the same way income tax brackets work.
We bill on requested vCPU rather than limits or actual usage because requests are what your scheduler reserves. Spot, on-demand, RI, and Savings Plan nodes all bill at the same per-vCPU rate — Kubeadapt's rate is independent of your cloud bill.
What counts as billable vCPU
The agent samples the requested vCPU on every running pod once per minute and averages over the billing month. The total that lands on your invoice is the average billable vCPU across the whole month, minus the always-free allowance.
What counts
- Pod CPU requests on Running pods, sampled per minute.
- All workload kinds — Deployments, StatefulSets, DaemonSets, Jobs, CronJobs, bare pods.
- Init containers during their actual run window.
- Sidecars and additional containers in a pod.
What doesn't
- CPU limits — only requests count.
- Actual usage — even idle workloads count if they request vCPU.
- System namespaces —
kube-system,kubeadapt,kube-public,kube-node-lease. - The agent itself — Kubeadapt's own pods are excluded automatically.
- Pending / Failed pods — only Running pods count.
- Memory and storage — billing is purely vCPU-based.
Worked example
A cluster averages 240 vCPU of requests across the month. The first 100 are free, the remaining 140 sit in tier 1 (0–500 billable) at $1.99 / vCPU.
(240 − 100) × $1.99 = $278.60 / month
A startup with the same 240 vCPU pays $84.00 / month on the flat $0.60 rate.
A larger cluster at 6,100 vCPU (6,000 billable) walks through all four tiers: $6,800 / month on Standard, $3,600 / month on Startups.
The 100 vCPU free allowance
Every Kubeadapt account gets 100 vCPU free, forever, with the full feature set — the same Cost Monitoring, Rightsizing, Best Practices, CLI, and API as paid plans.
- The allowance is account-wide, not per-cluster. Five clusters at 20 vCPU each (100 total) pays nothing.
- A cluster at 105 vCPU pays for the 5 above the allowance — 5 × $1.99 = $9.95 / month at tier 1 of the standard rate.
- Brief spikes are averaged. Two hours at 130 vCPU then back to 90 bills only the extra 30 vCPU for those two hours, at the tier rate that slice falls in.
- No credit card required to start. Add a payment method later if you grow past the allowance.
Plans
Same features on every tier. The plan you're on only changes the per-vCPU rate and the support response time.
Standard
Tiered: $1.99 → $0.60
Best for Most teams
- Self-serve sign-up at app.kubeadapt.io
- Best-effort support (business-day response)
- Monthly invoicing by default
Startups
$0.60 / vCPU / month
Best for < 3 yrs · < 30 staff · < $6M raised
- Flat rate, no tier math to track
- Self-attested at sign-up, audited annually
- Same support and features as Standard
Enterprise
Custom
Best for On-prem · air-gap · custom MSA
- Self-managed, on-prem, or air-gapped install
- Quarterly minimum, locked rate per contract
- Named support engineer, custom SLA, MSA, DPA
Outgrow startup eligibility (raised more than $6M, hit your 30th employee, or celebrated your third birthday)? You move to Standard at the start of the next billing month, never mid-month.
Billing cycle
Two cycles. Pick whichever matches how your finance team prefers to pay.
Default
Monthly
Charged on the 1st of the following month for the previous month's usage. Cycle resets each calendar month — partial months are pro-rated by the day.
Save 16.7%
2 months freeAnnual prepay
Pay 10 months upfront, get 12. Overage above your prepaid commit settles monthly at the same per-vCPU rate. Multi-year contracts available for enterprise.
Payment methods and currencies
We bill in USD by default. EUR is available on request. All payment processing is handled by Stripe; we never see or store your card details.
Credit card
Visa, Mastercard, Amex
ACH
US bank transfer
SEPA Direct Debit
EU/UK bank transfer
Wire transfer
On request, $1k+ / mo
Net-30 invoicing is automatic on annual contracts and on monthly accounts billing more than $1,000. Need different terms (Net-45, Net-60, PO matching)? Email billing@kubeadapt.io.
Invoices and receipts
Every invoice is issued PDF, hosted on Stripe, and emailed to your billing contact within 24 hours of close. Past invoices live in the dashboard under Settings → Billing → Invoices.
Invoice line items
Average billable vCPU, free allowance applied, per-cluster breakdown, taxes if applicable.
CSV export
Daily billable vCPU per cluster, downloadable from the dashboard for ledger reconciliation.
Need a custom billing email, PO number on the invoice, or reverse-charge VAT? Set it under Settings → Billing → Tax & Address — or email us with your VAT ID and we'll do it for you.
Failed payments and grace period
Card declined? Bank holiday? Your invoice retries automatically and we email you first. Nothing breaks immediately.
- Day 0Payment fails. Stripe retries on day 3, 5, and 7. The dashboard shows a banner and emails go to the billing contact.
- Day 1–14Grace period. Everything keeps working — agents collect data, recommendations refresh, the dashboard is fully interactive.
- Day 15+The dashboard goes read-only. You can still view historical data and export CSV. Recommendations stop refreshing. The agent in your cluster keeps running.
- ResolvedPay the outstanding invoice and full functionality returns within minutes. No data is lost during the grace period or after.
Cancellation and your data
Cancel any time from Settings → Billing → Cancel subscription. No call required, no retention pitch.
- Monthly plans end at the next billing date. You're billed for usage up to that point and nothing more.
- Annual prepay stops auto-renewing. You keep paid access until the end of the prepaid term. We don't pro-rate refunds for unused months.
- Data is retained for 5 years after cancellation, then deleted permanently. Export anything you need (CSV, API) before that window closes.
- The agent stops on uninstall, not on cancel. See the uninstall steps to remove the Helm release from your cluster.
Common questions
For pricing-model questions (vCPU mechanic, startup eligibility), see the pricing FAQ.
Do I get charged twice if a pod is rescheduled?
No. We bill on average requested vCPU per minute. A pod rescheduled from one node to another in the same minute counts once.
What if I install Kubeadapt mid-month?
Your first invoice covers from install date to month-end and is pro-rated. The 100 vCPU free allowance applies in full from day one — it's not pro-rated.
Are spot or preemptible nodes billed differently?
No. Kubeadapt's billing depends on pod requests, not node type. A 100m request bills the same on a spot node as on an on-demand node.
Can I get a custom invoice format or PO number?
Yes. Set the PO number, billing email, and any required reference fields under Settings → Billing, or email billing@kubeadapt.io.
Do you offer multi-org / parent billing?
Yes. Multiple Kubeadapt orgs can roll up to a single billing account. Useful for enterprises with separate dev/staging/prod tenants. Email sales to set it up.
Need help?
We answer billing questions same business day.
Invoice question, custom contract, refund request, or just want the math worked out for your environment — email us. Real engineers reply.