FluxBilling
Feature · VPS

Your KVM machines, sold as virtual servers.

Register a machine, describe the sizes you want to sell, attach one to a product. From then on a paid order builds a real guest on your hardware — addresses, disks, console and invoice included — without anyone touching a hypervisor by hand.

KVM · qemu + libvirtoutbound-only agentlive migrationIPv4 + IPv6 from IPAM
client portal · vps
FluxBilling client portal — a running virtual server showing hostname, operating system, location, its allocated IPv4 address, and live CPU, RAM, disk and bandwidth usage
What it costs

€6 per node, per month. That is the whole meter.

A node licence covers one machine you run the platform’s agent on. What that machine does with itself is not metered: run virtual servers on it, run game containers on it, or run both side by side — it is one licence either way. Guests are not counted, cores are not counted, and there is no published ceiling on how many nodes you buy.

Node licences are added to any tier at order time, or to a licence you already hold. They are separate from the bare-metal server meter, which counts dedicated servers you sell rather than machines you own.

Node licence
Price
€6 / node / month
Covers
Virtual machines and game containers
Counted per
Machine running the agent
Quantity limit
None published
Guests per node
Not metered
Model

Six pieces, and nothing hidden behind them

The panel’s own screens call a registered machine a cluster, and use “cluster” and “node” interchangeably. One machine is one cluster — there is no such thing as a cluster containing several machines.

PieceWhat it is
NodeOne physical hypervisor machine, registered with the panel.
AgentThe small program installed on that machine. It is the only thing that ever talks to the hypervisor: the panel writes an instruction, the agent collects it over an outbound connection and carries it out. A machine with no agent cannot be managed.
GroupA named set of nodes. A group decides which machine a new guest lands on, and it is the boundary inside which a running machine can be moved live from one node to another.
Storage poolA place on one node where disks are created — a volume group, a directory of disk images, or a storage dataset. Each node reports the pools it has; you choose which ones you sell from.
PlanThe size you sell: cores, memory, disks, addresses, which operating systems it offers, and which locations, groups or nodes may host it. A plan is edited on the product that sells it.
Virtual machineOne running guest, built from a plan on a node. It may be linked to a billing service, or unlinked — built for yourself, or adopted from another panel.

Addresses come from IPAM. You can reserve part of your address space for virtual machines and point it at particular nodes or groups, so a physical server can never be handed an address you were keeping for a guest.

Onboarding

You supply the machine. The panel installs itself onto it.

Give the panel an IP, an SSH port and a root login. It installs the agent, the KVM stack and a time-synchronisation service, then waits for the machine’s first check-in and reads its inventory — bridges, storage pools, CPU and memory. Roughly 3 to 15 minutes, and you can close the window while it runs.

SSH is the only inbound access the install needs. Everything after it is pulled by the agent over HTTPS, so a node behind NAT works with no inbound rule at all.

A machine that already runs another virtualisation panel is brought in through VPS Panel Migration instead, which adopts its running guests rather than starting from nothing.

What the machine needs firstprerequisites
  • 64-bit LinuxHardware virtualisation available and enabled in firmware.
  • qemu-kvm + libvirtThe panel can install these for you; install them first if you add the agent yourself.
  • Inbound SSHRoot login by password or private key — for the install only.
  • Outbound HTTPSEverything after the install is pulled by the agent.
  • Somewhere for disksA volume group, a directory, or a storage dataset. The panel reads what exists; it never converts your storage.
  • A network bridgeNodes report the bridges they have; one of them carries guest interfaces.
  • A correct clockA node whose clock drifts badly refuses to run guest work.
Operations

“Is anything wrong, and how much room is left”

The module opens on a fleet view that answers exactly that. Detected conditions are listed as Critical, Warning or Info, each linking to the node or guest it concerns.

  • A disk reporting itself as failing, or running low on write life
  • Host memory pressure on a node
  • Memory sold beyond the physical amount
  • A storage pool filling up
  • A node that has stopped sending samples
  • An agent that has gone offline
  • A node frozen or unreachable
  • Failed agent jobs in the last 24 hours
  • Guests sitting in an error state

Sold and used are never mixed. Sold is what has been committed to customers; used is what the machines are actually doing. A node at 40% used and 200% sold is what overselling looks like, and is fine. Where nothing has reported a figure the panel prints Not reported rather than a zero — an unmeasured machine must never read as an idle one.

Client portal

The customer never opens a ticket to reboot

PowerStart, stop, restart. Every action is a job with an audit trail, not a fire-and-forget button.
ConsoleBrowser VNC to the guest, relayed through the panel — no port exposed on the node.
ReinstallRebuild from any operating system the plan offers, with a fresh root credential.
Backups & graphsSnapshots within the plan’s allowance, plus CPU, memory, disk and network history.

Suspension, termination and renewal are the same lifecycle every other service on the platform uses, so an unpaid virtual server behaves like an unpaid dedicated server — see Billing and Provisioning.

Get started

Try it on your own data. Refund inside 14 days if it’s not the fit.

Pick a tier and provision a tenant in under two minutes — isolated K3s namespace, your own database, the full product. If FluxBilling isn’t the right fit inside 14 days, open a ticket and we’ll refund the subscription. No sales call, no qualification gate.

14-day refund
Standard policy
  1. 01.
    Pick a tier
    Lite from €4.95/mo. Upgrade later, no migration.
    < 1 min
  2. 02.
    Provision the tenant
    Isolated K3s namespace + your own PostgreSQL database. Full product, your data.
    < 2 min
  3. 03.
    Refund inside 14 days
    Not the fit? Open a ticket within 14 days and we refund the subscription. No questions, no qualification gate.
    d0 — d14
14-day refund · cancel any time
Start →