# Reserved Instances

## What are Reserved Instances?

Reserved Instances are a pricing option offered by AWS that provide a significant discount on EC2 usage in exchange for committing to a specific instance type, region, and term length. Unlike Savings Plans, the commitment is to a particular configuration rather than a dollar amount.

The discount is the reward for predictability. The more specific your commitment, the deeper the rate reduction.

## How Reserved Instances work

When you purchase a Reserved Instance, you are reserving capacity for a defined configuration: instance type, region, operating system, and tenancy. AWS applies a discounted rate to matching usage for the duration of the term, either one or three years.

There are three payment options:

- **All upfront:** The full amount is paid at purchase, which gives you the deepest discount over the term.
- **Partial upfront:** A portion is paid at purchase and the remainder is billed monthly, splitting the cost while still offering a meaningful discount.
- **No upfront:** The commitment is billed monthly throughout the term, which is the easiest on cash flow but carries the smallest discount of the three.

The trade-off is between cash flow and savings. Paying more upfront means a lower effective hourly rate over the term.

## Types of Reserved Instances

**Standard Reserved Instances** are locked to a specific instance family, region, and OS. They offer the deepest discounts and work best for stable, predictable workloads where the configuration is unlikely to change.

**Convertible Reserved Instances** can be exchanged for a different configuration during the term, which gives you more flexibility if your infrastructure needs shift. The discount is slightly lower than Standard Reserved Instances in exchange for that flexibility.

## Reserved Instances in action

A company runs a production web application on a fleet of m5.xlarge EC2 instances in us-east-1. The fleet has been stable for over a year and is expected to remain so.

|  | On-demand | Reserved Instance (1 year, all upfront) |
|---|---|---|
| Hourly rate | $0.192 | $0.112 |
| Monthly cost (10 instances) | $1,382 | $806 |
| Annual saving |  | ~$6,900 |

The application runs the same as before—at roughly half the cost.

## Reserved Instances vs. Savings Plans

Both offer discounts in exchange for a commitment, but they suit different situations:

|  | Reserved Instances | Savings Plans |
|---|---|---|
| Commitment type | Specific instance, region, OS | Hourly spend amount |
| Flexibility | Lower | Higher |
| Discount depth | Up to 72% | Up to 66% |
| Best for | Stable, well-defined workloads | Variable or evolving workloads |

Reserved Instances offer the deeper discount, but only if the commitment matches actual usage closely. If the instance type or region changes during the term, you may end up with unused Reserved Instances that are still being billed.

## What to watch out for

Reserved Instances are straightforward in principle, but a few things tend to catch teams off guard when purchasing or managing them:

- Purchasing Reserved Instances before rightsizing workloads locks in a committed rate on oversized instances. Right-size first, then commit.
- Unused Reserved Instances still incur charges. Monitoring Reserved Instances utilization regularly helps catch coverage gaps early.
- Standard Reserved Instances cannot be modified once purchased. If your workload moves regions or changes instance family, the Reserved Instance becomes stranded.
- In multi-account AWS Organizations, unused Reserved Instances in one account can be shared across others, which reduces waste but requires visibility across the organization.

## Reserved Instances coverage vs. utilization

These are two distinct metrics that are easy to confuse:

- Reserved Instances coverage measures what percentage of your eligible on-demand usage is covered by Reserved Instances. Low coverage means you are leaving discount potential on the table.
- Reserved Instances utilization measures how much of your purchased Reserved Instance capacity is actually being used. Low utilization means you are paying for commitment you are not using.

A team can have high coverage but low utilization if they over-purchased, or high utilization but low coverage if they under-purchased. Both numbers together give a more complete picture of how well your commitments are working.

## Azure and GCP equivalents of Reserved Instances

Reserved Instances is an AWS product. The equivalent offerings on other platforms include:

- Azure offers Azure Reservations for virtual machines, databases, and other services
- GCP offers Committed Use Discounts for Compute Engine workloads

The commitment mechanics differ across providers, so each should be evaluated on its own terms; don't make the mistake of assuming they all work the same way.

## Explore related content

### Savings Plans

[AWS Savings Plans offer discounts on compute usage in exchange for a committed hourly spend. Learn how they work, the difference between plan types, and how they compare to Reserved Instances.](https://www.manageengine.com/cloudspend/finops-glossary/saving-plans.html)

### Azure Reservations

[Azure Reservations offer discounts on virtual machines, databases, and other Azure services in exchange for a one or three year commitment. Learn how they work and what to get right before purchasing.](https://www.manageengine.com/cloudspend/finops-glossary/azure-reservations.html)

### GCP Committed Use Discounts

[GCP Committed Use Discounts reduce Compute Engine costs in exchange for a one or three year commitment. Learn how resource-based and spend-based CUDs work and what to get right before purchasing.](https://www.manageengine.com/cloudspend/finops-glossary/gcp-committed-use-discounts.html)