Reserved Instances
Reserved Instances are a cloud pricing model in which a customer commits to using a specific amount of compute capacity for a fixed term — typically one or three years — in exchange for a significantly lower price than standard on-demand…
Definition
Reserved Instances are a cloud pricing model in which a customer commits to using a specific amount of compute capacity for a fixed term — typically one or three years — in exchange for a significantly lower price than standard on-demand rates.
Overview
Reserved Instances (RIs), most closely associated with AWS but with equivalents on every major provider (Azure Reserved VM Instances, Google Cloud Committed Use Discounts), trade flexibility for savings. Instead of paying the full on-demand hourly rate for compute, a customer commits upfront to running a certain instance type (or, with more flexible variants, a certain dollar amount of usage) for one or three years, and receives a substantial discount — commonly 30-70% off on-demand pricing — in return. This makes RIs a good fit for predictable, steady-state workloads: a database that runs 24/7, a baseline fleet of web servers that's always needed regardless of traffic, or any capacity a team is confident it will use continuously for the length of the commitment. It's a poor fit for spiky, unpredictable, or short-lived workloads, which are better served by on-demand pricing or, for fault-tolerant batch work, Spot Instances. Payment options usually range from all-upfront (cheapest), to partial-upfront, to no-upfront (a bit more expensive but with no large initial cash outlay), letting organizations balance savings against cash flow. RI purchasing decisions are a central part of FinOps practice, since committing too aggressively locks in cost even if usage drops, while committing too conservatively leaves savings on the table; tools like AWS Cost Explorer analyze historical usage and generate RI purchase recommendations to help strike that balance. Many organizations now favor Savings Plans, a more flexible successor to traditional RIs that commits to a dollar amount of spend rather than a specific instance type.
Key Concepts
- One- or three-year capacity commitment in exchange for discounted pricing
- Discounts typically range from roughly 30% to 70% off on-demand rates
- All-upfront, partial-upfront, and no-upfront payment options
- Standard RIs lock in a specific instance type; convertible RIs allow some flexibility to change it
- Best suited for predictable, steady-state, always-on workloads
- Complemented by Savings Plans, a more flexible spend-based commitment model
Use Cases
Frequently Asked Questions
From the Blog
How to rightsize cloud compute instances
Resize instances without causing incidents: which utilisation percentiles to measure, how long to observe, sizing to peak versus average, and staging the change.
Read More Cloud & CybersecurityChoosing between EC2, containers and Lambda
Pick the right AWS compute model: how request duration, burstiness, cold starts, packaging and operational load push a workload toward instances, containers or functions.
Read More ProgrammingHow to fix layout shift caused by images, fonts and ads
Every layout shift traces back to space that was not reserved, so the fix is always reservation. Identify the shifting element from the layout shift entries first, since the visible symptom is often not the element that moved, then apply the specific remedy for media, fonts and late-arriving content.
Read More