Guides

Cloud billing concepts explained: hourly, monthly, reserved and committed use

The vocabulary of cloud billing, from hourly rates and monthly caps to reserved capacity, savings plans, spot pricing and prepaid credits.

Updated September 25, 2026

Cloud billing concepts explained: hourly, monthly, reserved and committed use

Cloud pricing pages assume you already speak the language. Hourly, per-second, monthly cap, reserved, committed use, savings plan, spot, prepaid credit: each term describes a different agreement about when you pay and how much risk you carry. This guide defines the terms and shows where they appear on two very different platforms, Kamatera and AWS.

Usage-based billing: hourly and per-second

The default model on most platforms is consumption. A resource accrues charges for as long as it exists, measured in hours, minutes or seconds. On AWS, EC2 instances are billed by the second, storage by the gigabyte-month and data transfer by the gigabyte. Kamatera lets you choose hourly billing when you create a server, which suits short-lived or variable workloads.

The important word is exists. A stopped virtual machine may still incur storage charges for its disk, and an unattached volume or idle IP address keeps billing even though nothing is running on it. Usage-based billing rewards teams that delete what they no longer need.

Monthly rates and monthly caps

Some providers offer a flat monthly price, either as an alternative to hourly billing or as a ceiling on it. Kamatera offers hourly or monthly billing on the same configuration, so you pick whichever suits the expected lifetime of the server. Other developer-focused platforms accrue hourly up to a monthly cap, which gives predictable maximums without penalizing short use.

Monthly billing is simplest to budget for and works well for servers that will run continuously. Hourly billing wins when environments are created and destroyed frequently, such as test systems or build agents.

Component pricing

Kamatera’s model is worth understanding on its own because it is unusually transparent. You select a CPU type (Availability, General Purpose, Dedicated or Burstable), then independently choose the number of CPUs, memory, storage and traffic, and finally choose hourly or monthly billing. The price is the sum of the components. That makes it easy to see what a change in memory or storage will cost before you make it. The Buy Kamatera Account page lists a “Kamatera 5 Limit with $100 Credit Account”, and the Kamatera hourly billing guide explains how the limit and the credit interact.

Reserved capacity, savings plans and committed use

Hyperscale platforms offer discounts in exchange for commitment. The names differ but the idea is the same: promise to use a certain amount of compute for one or three years and pay less per unit.

  • Reserved instances commit to a specific instance type in a specific region.
  • Savings plans commit to a spending level per hour and apply across instance families and, in some cases, regions. AWS offers Compute Savings Plans and EC2 Instance Savings Plans for steady workloads.
  • Committed-use discounts is the term some other providers use for a similar spend-based or resource-based commitment.

Commitments only pay off when usage is genuinely predictable. A new account rarely qualifies on day one, which is why the cost control guide for new accounts recommends measuring for a few billing cycles before committing.

Spot and interruptible pricing

Spot pricing sells spare capacity at a discount with the condition that the provider can reclaim it at short notice. AWS Spot Instances suit interruptible jobs such as batch processing, rendering or some training workloads. The saving is real but the workload must tolerate being stopped, so spot is a fit for architecture, not a general discount.

Prepaid balances and credits

A credit is a balance that offsets eligible charges before the payment method is used. On AWS, credits are applied automatically against eligible charges, and the marketplace lists options such as “AWS $1K Credit Account” and “AWS $5K Credit Account” on the Buy AWS Account page. Credits can carry service restrictions and expiry dates, so confirm both before relying on one. The how cloud credits work article compares credit behavior across providers.

Putting the vocabulary together

Model You pay for Best fit
Hourly or per-second Time the resource exists Temporary and variable workloads
Monthly rate or cap A fixed month, sometimes as a ceiling on hourly Always-on servers
Reserved or committed A one- or three-year promise at a discount Stable, measured production usage
Spot Spare capacity that can be reclaimed Interruptible jobs
Credit or prepaid Nothing until the balance is consumed Early experimentation with a defined budget

Most accounts end up using several of these at once: a credit for the first months, hourly billing for development, and a commitment for the production core once it is measured. For a side-by-side view of listed options by billing model, see pay-as-you-go versus credit accounts.

Questions on this topic

Is hourly billing always cheaper than monthly?

No. Hourly billing is cheaper when a server runs for part of a month and is deleted afterwards. For a server that runs continuously, a monthly rate or a monthly cap is usually the same or lower and easier to budget. Kamatera lets you pick either model per server for exactly this reason.

When should a new account consider reserved or committed pricing?

Only after usage has been measured across a few billing cycles. Commitments discount a predictable baseline, and a new account rarely knows its baseline yet. Start on usage-based billing, use any credit as a buffer, then commit the stable production core once the pattern is clear.

Do credits change how the underlying billing model works?

No. A credit is a balance that offsets eligible charges; the resources still accrue at hourly, per-second or monthly rates underneath. When the balance is consumed or expires, the same charges are billed to the payment method. Check any service restrictions and expiry before relying on a credit.

More from the Blog

All articles