A reserved instance purchase checklist is filled in by the cloud owner before every commitment purchase, whether reserved instance, savings plan, Azure reservation or committed use discount, and the approving role signs it; the record holds the baseline usage, the coverage target, term and payment option, the programme's exchange and refund rules, the approval and the review date, one record per purchase. The decision it records: buy, reduce, defer or reject, with reasoning. Which programme fits which load is explained on choosing between commitment types; how purchases become a strategy on commitment strategy. This page is the record.
The template
Commitment purchase decision record <id>, date <date>, requested by <cloud owner>, approved by <role>
1. Subject
| Field | Value |
|---|---|
| Provider and programme | <AWS Compute Savings Plan, EC2 Instance Savings Plan, Standard or Convertible Reserved Instance; Azure reservation or savings plan; Google Cloud resource-based or spend-based CUD> |
| Size | <amount per hour, number of instances, vCPU and GB, or amount per month> |
| Scope | <account or organisation>, <region>, <instance family, operating system, tenancy> |
| Term | <1 or 3 years>, start <date>, end <date> |
| Payment option | <all upfront, partial upfront, monthly> |
| Discount per the provider | <percent against on-demand, price list of date, link> |
2. Baseline usage
| Figure | Value | Period | Source |
|---|---|---|---|
| Usage of the subject per hour, minimum | <value> | last <90> days | export, filter <usage type, meter or SKU> |
| Usage per hour, mean | <value> | last <90> days | export |
| Of which already covered by commitments | <percent> | last month | coverage report |
| Utilisation of existing commitments | <percent> | last month | coverage report |
| The provider's recommendation | <size, amount> | parameters <days, term, payment> | <Cost Explorer, Azure Advisor, Recommender> |
| Planned changes that lower the load | <rightsizing, move to ARM, shutdown, end date of a project> | within the term | <roadmap, team lead> |
3. Arithmetic
| Item | Value |
|---|---|
| Cost of the commitment over the term | <amount> |
| On-demand cost of the same usage | <amount> |
| Saving over the term | <amount>, <percent> |
| Break-even utilisation | <percent of hours from which the commitment beats on-demand> |
Unused amount at <20> percent less load | <amount per month> |
| Coverage after the purchase | <percent>, target <percent> |
4. Programme rules
| Rule | For this purchase |
|---|---|
| Change or exchange during the term | <per the provider: possible, limited, not possible> |
| Cancellation or refund | <per the provider: not possible, up to an amount per year, through a marketplace> |
| Flexibility across size, family, region | <per the provider> |
| Sharing across accounts or subscriptions | <scope shared: yes or no> |
| Automatic renewal | <on, off>, checked on <date> |
| Appears in the billing data as | <columns or meters that carry the fee and the coverage> |
5. Approval and review
| Field | Value |
|---|---|
| Approved by | <role per the RACI; from an amount of term value, management> |
| Decision | <buy, reduce to size, defer until date, reject>, reasoning <text> |
| Budget line | <sheet 3 of the budget, row> |
| Utilisation review | <30> days after purchase, then monthly, owner <role> |
| Reminder before expiry | <90> days before <end date>, to <role> |
Checklist before signing
- Baseline from at least
<90>days, not from last week or a peak month - Size below the minimum of the baseline, not below the mean
- Existing commitments with utilisation below
<percent>explained or exchanged first - Planned changes within the term checked with the team leads
- Instance generation current; no change of family or move to ARM planned
- Payment option agreed with finance: liquidity and the posting of the prepayment
- Scope set to the organisation where several accounts are to use the coverage
- Review date and expiry reminder entered, record linked in the budget
How to fill it in
Subject: the size in the unit the provider bills, amount per hour for savings plans and spend-based CUDs, instances for reservations; the programmes per provider are on AWS Reserved Instances, AWS Savings Plans, Azure reservations, Azure savings plans and committed use discounts.
Baseline: the minimum per hour is the figure the commitment must not exceed; the mean shows what stays on-demand. The provider's recommendation has parameters you write down, because they determine the result. Arithmetic: the break-even utilisation follows from discount and payment option; how coverage and utilisation are measured is on commitment coverage and utilisation.
Rules: every row from the provider's current documentation, with a date, because exchange and refund rules change; the state per programme is summarised on commitment exchange, cancellation and expiry. How the fee appears in the exports is on list, billed, effective and amortised cost.
Approval: the threshold from which management signs is in the RACI; the budget line in the budget template.
How to adapt it
One team: requester and originator are the same; section 2 is filled in together with the team lead. Several teams: coverage is bought at organisation level, and section 2 sums the baseline across the accounts, with one row per team whose end dates have been checked.
One provider: section 4 carries the rules once and is only dated at each purchase. Three providers: one record per provider and programme, because exchange, refund and flexibility differ; coverage is measured per provider.
Startup: one-year terms only, monthly payment and a size below half the minimum until two quarters of baseline exist. Finance-led: payment option and posting agreed with finance before the arithmetic in section 3, and the record is filed with the provider's invoice.
Sources
Read on 2026-10-21. FinOps Foundation, Rate Optimization (CC BY 4.0), What are Savings Plans?, understanding your Savings Plans recommendation calculations, Reserved Instances for Amazon EC2, decide between a savings plan and a reservation in Azure, exchanges and refunds for Azure reservations, Azure savings plan recommendations, committed use discounts, recommendations for committed use discounts.