New See exactly what you're overpaying AWS in under 60 seconds. Try the Calculator for free

How to Save on RDS Reserved Instances

Learn how to right-size RDS workloads, compare Reserved Instances with Database Savings Plans, and reduce commitment risk.
Updated September 1, 2026
18 min read
How to Save on RDS Reserved Instances
In this article
Key takeaways
1
RDS Reserved Instances can save up to 69% versus On-Demand rates for steady-state workloads, but the actual discount depends on the current offering and configuration.
2
Right-size first, then reserve only the stable baseline. A discount on an oversized database can lock unnecessary spend into the commitment.
3
Size flexibility applies to Aurora, MySQL, MariaDB, PostgreSQL, Db2, and Oracle BYOL. SQL Server and Oracle License Included do not receive RDS size flexibility.
4
After purchase, monitor utilization. RDS Reserved Instances cannot be cancelled, transferred, or sold.
Amazon RDS Reserved Instances can significantly reduce database costs when your workloads have predictable, long-term usage. By committing to a specific level of RDS usage, you can pay less than standard On-Demand rates.

The key is knowing which workloads to reserve, how to choose the right term and payment option, and how to avoid paying for unused capacity.

Short answer

RDS Reserved Instances are best for predictable database usage that is likely to remain stable for the commitment term. AWS offers one-year and three-year terms and advertises savings of up to 69% over On-Demand rates.

Before buying, right-size the database, check matching and lifecycle rules, compare the RI’s effective hourly cost with On-Demand pricing, and evaluate whether a Database Savings Plan provides better flexibility.

What is an RDS Reserved Instance?

An RDS Reserved Instance is a billing discount, not a separate database. AWS automatically applies the discounted rate when eligible usage matches attributes such as Region, database engine, DB instance type, edition, and license model. See the AWS Reserved DB instance documentation for the current matching and billing rules.

Deleting a database does not cancel the reservation. The commitment continues for the purchased term.

For an engine-by-engine comparison, see our RDS Reserved Instance pricing guide.

How much can RDS RIs save?

AWS states that RDS Reserved Instances can save up to 69% over On-Demand rates when used in steady state. Your actual discount depends on the offering available for your engine, class, Region, term, and payment option.

Reserved pricing discounts eligible DB instance usage. Storage, backups, data transfer, and RDS Extended Support remain separate from the RI discount. I/O charges can also remain separate where the selected configuration bills for I/O, but Aurora I/O-Optimized has no additional charges for read and write I/O operations.

How to compare the actual offers

AWS recommends comparing the On-Demand hourly rate with the RI’s effective hourly rate after amortizing any upfront payment over the term.

Effective RI hourly cost = Upfront fee ÷ total term hours + recurring hourly fee

Savings per hour = On-Demand hourly rate − effective RI hourly cost

Savings percentage = Savings per hour ÷ On-Demand hourly rate × 100

Hypothetical example: Assume the matching On-Demand rate is $1.00 per hour. A one-year Partial Upfront RI requires $1,752 upfront plus $0.40 per hour. Amortizing $1,752 across 8,760 hours adds $0.20 per hour, making the effective RI rate $0.60. The modeled saving is 40%.
Compare offers using the same engine, Region, DB instance class, deployment configuration, term, and payment option.

Which engines get size flexibility?

RDS size flexibility lets a reservation apply proportionally across sizes within the same instance class type. AWS currently lists Aurora, MySQL, MariaDB, PostgreSQL, Db2, and Oracle BYOL as size-flexible. It does not apply to SQL Server or Oracle License Included.

AWS uses normalized units. A Single-AZ large equals 4 units, xlarge equals 8, and 2xlarge equals 16. Size-flexible benefits can also apply across Single-AZ and Multi-AZ configurations when the other matching requirements are met. The AWS Reserved DB instance documentation explains the current normalization and flexibility rules.

For a deeper example, see our RDS MySQL RI size flexibility guide.

Six steps to maximize RDS RI savings

1. Right-size before you reserve

Review a representative workload period across CPU, memory pressure, connections, I/O, throughput, and latency. Validate the smallest safe class through normal production peaks. Reserving an oversized database discounts the wrong baseline.

Our RDS MySQL right-sizing guide covers the workflow in more detail.

2. Check Extended Support exposure

Verify the engine lifecycle before buying a multi-year RI. RDS Extended Support is charged separately, and RI discounts do not reduce those charges. MySQL 5.7, for example, remains in Extended Support, with support extended through June 30, 2029.

3. Audit the matching attributes

Inventory Region, engine, instance class type, edition, license model, and deployment configuration. For size-flexible engines, calculate normalized-unit coverage.

If you use AWS Organizations with consolidated billing, assess eligible organization-wide usage. AWS can apply the hourly benefit of a reserved DB instance across accounts in the consolidated billing family, so the safe baseline may be larger than usage in the purchasing account alone.

4. Choose the term and payment option

RDS Reserved Instances support All Upfront, Partial Upfront, and No Upfront payment options. All Upfront and Partial Upfront can use one-year or three-year terms. No Upfront is available only for one year and may require AWS approval.

Choose a three-year term only when the underlying database baseline has a credible three-year horizon. See our one-year versus three-year RDS RI guide for term selection.

5. Compare Database Savings Plans

Database Savings Plans use a one-year, no-upfront hourly commitment. The hourly commitment is entered at the Savings Plans rate, not the equivalent On-Demand spend.

For eligible provisioned RDS instance usage, Database Savings Plans can provide up to 20% savings; eligible serverless database usage can save up to 35%. They can also follow supported changes across database services, engines, instance families, sizes, Availability Zones, and Regions. See the AWS Database Savings Plans documentation for current coverage.
Situation Likely starting point Why
Stable RDS baseline RDS Reserved Instance Potentially deeper discount
Engine, family, or Region may change Database Savings Plan More flexible eligible coverage
Short-lived or uncertain workload On-Demand Avoid unused commitment
Existing RI already covers usage Keep RI in the model Avoid double-counting coverage
Decision example: A team runs RDS for Oracle today but expects to modernize to Aurora PostgreSQL or move part of the workload to DynamoDB. An RDS RI may offer a stronger rate on today’s Oracle baseline, while a Database Savings Plan can be more resilient because eligible benefits can continue across supported database services and configuration changes.

Use our Database Savings Plans guide for the broader commitment model.

6. Purchase, then monitor

Purchase only the stable baseline. Review utilization and coverage regularly so deleted, migrated, or resized databases do not leave paid commitment unused.

Stranded-spend risks to check

Do not reserve an oversized database just to maximize coverage.

Do not assume RDS has EC2-style Convertible RIs. RDS RIs cannot be exchanged like EC2 Convertible RIs.

Do not choose a long term before checking engine lifecycle, migration plans, and Database Savings Plan alternatives.

Pre-purchase guide

Confirm the database has been right-sized using representative workload data.

Check the engine version for Extended Support exposure.

Verify Region, engine, class type, edition, license model, and size-flexibility rules.

If consolidated billing is enabled, include eligible organization-wide usage.

Compare the RDS RI effective hourly cost with On-Demand pricing.

Compare the RI with an appropriate Database Savings Plan scenario.

Confirm the stable baseline should remain for the selected term.

Review upfront payment and commitment risk with finance.

How to purchase RDS Reserved Instances

In the Amazon RDS console, open Reserved instances, choose Purchase reserved DB instance, select the product description, DB instance class, Multi-AZ option, term, and offering type, then review current pricing before submitting. AWS documents the current steps in its reserved DB instance purchase guidance.

For CLI-based purchasing, first list available offerings, then purchase using the returned offering ID:

Shell

aws rds describe-reserved-db-instances-offerings \
  --db-instance-class db.r8g.xlarge
aws rds purchase-reserved-db-instances-offering \
  --reserved-db-instances-offering-id <offering-id> \
  --reserved-db-instance-id <reservation-id> \
  --db-instance-count 1
Verify the current offering, term, payment option, and effective hourly cost before purchasing.

How to monitor after purchase

Use AWS Cost Explorer reservation reports to review active RDS utilization. Investigate persistent underuse by checking for deleted databases, family or Region changes, engine migrations, or exact-size mismatches on engines without size flexibility.

If a covered database is deleted, another compatible DB instance can receive the benefit during the remaining term. Because the RI cannot be cancelled, utilization monitoring is part of commitment management.

Where Usage.ai fits

At Usage.ai, we analyze usage, surface commitment recommendations, and, after approval, can purchase commitments on your behalf using scoped cloud-provider access. We charge a percentage of realized savings.

With our Flex Insured Commitments program, teams can capture up to 69% savings available through Amazon RDS Reserved Instances without taking on the long-term commitment risk. If a commitment no longer delivers savings compared with equivalent On-Demand usage, Usage.ai provides cashback protection to help cover the difference.

The result is a managed RDS commitment strategy that combines AWS database discounts with automation, flexibility, and protection against commitment risk.

Conclusion

RDS Reserved Instances can materially reduce steady database instance costs, but the best commitment is not automatically the one with the largest advertised discount. Right-size first, calculate the effective hourly rate, account for organization-wide eligible usage, compare RIs with Database Savings Plans, and commit only to a baseline you expect to keep.
RDS commitment planning
Reduce RDS commitment risk

See your RDS savings and reduce commitment risk.

Frequently asked questions

Can I cancel or convert an RDS Reserved Instance?

No. AWS states that RDS Reserved Instances cannot be cancelled, transferred, or sold. RDS also does not offer the EC2 Convertible RI model. Eligible engines instead rely on RDS size-flexibility rules for some configuration changes.

Can one RDS RI cover several instance sizes?

Yes, for size-flexible engines. AWS uses normalized units within the same eligible Region, database engine, and instance class type. A reservation can fully cover smaller instances or partially cover a larger instance.

Can an RDS RI benefit another AWS account?

Yes, under consolidated billing. AWS states that accounts in an organization can receive the hourly cost benefit of reserved DB instances purchased by another account. Review the billing configuration before sizing commitments.

How are Database Savings Plans different from RDS RIs?

Database Savings Plans are one-year, no-upfront hourly commitments that can follow eligible database usage across supported services and configurations. RDS RIs are tied more closely to RDS matching attributes.

Do RDS RIs discount Extended Support?

No. RDS Extended Support charges are separate from DB instance charges, so RI discounts do not reduce them. Upgrade to a supported major engine version to eliminate Extended Support charges.

Share
Facebook
X
LinkedIn
Reddit
Cut cloud cost with automation
Latest from our blogs