AWS Database Savings Plans: The Complete 2026 Guide

Databases often account for 30-40% of total AWS spending. Yet many organizations overpay because they don't understand their savings options.

AWS Database Savings Plans changed that. They offer flexible, commitment-based discounts of up to 35% across nine managed database services.

This guide explains how Database Savings Plans work, what they cover, and when to use them instead of Reserved Instances.

Key Takeaways

• Save up to 35% on Aurora Serverless and up to 20% on RDS instance-based databases
• One plan covers nine services: Aurora, RDS, DynamoDB, DocumentDB, Neptune, ElastiCache (Valkey), Keyspaces, Timestream, and DMS
• Flexibility built-in: Discounts apply across instance sizes, families, engines, and regions
• First serverless discounts: Aurora Serverless v2 and DynamoDB on-demand now qualify
• Limitation: Older instance families (m5, r5, t3, t4g) require Reserved Instances instead

What Are AWS Database Savings Plans?

AWS Database Savings Plans are commitment-based discounts where you agree to spend a fixed dollar amount per hour on eligible database services in exchange for lower rates.

Here's the simple version:

1. You commit to spending $X per hour on databases (example: $10/hour)
2. AWS gives you discounted rates on all eligible usage
3. The commitment lasts one year with no upfront payment required

The key difference from Reserved Instances? Flexibility.

Your savings apply automatically across:

• Any instance size (db.r6g.large or db.r6g.4xlarge)
• Any instance family (r6g, r7g, x2g)
• Any database engine (MySQL, PostgreSQL, Oracle)
• Any AWS region

Change your database architecture mid-year? Your savings plan still applies.

Which AWS Database Services Are Covered?

Database Savings Plans cover nine AWS managed database services:

Important: ElastiCache for Redis and Memcached are not eligible. These still require Reserved Instances for discounts.

How Much Can You Save? (Discount Rates)

Savings vary by database type and configuration. Here are the typical discount percentages:

*Based on $36,000 annual database spend before discounts.

Why Serverless Discounts Matter

Before Database Savings Plans, serverless databases had zero commitment discount options. You paid full on-demand rates.

Now, Aurora Serverless v2 users can save up to 35%. DynamoDB on-demand users save around 18%.

This is significant for organizations moving toward serverless architectures.

5 Key Benefits of Database Savings Plans

1. Flexibility Across Your Entire Database Estate

Reserved Instances lock you into specific configurations. Database Savings Plans don't.

Your commitment applies automatically to whatever databases you run. Migrate from RDS MySQL to Aurora PostgreSQL? Still covered. Move workloads to a different region? Still covered.

2. First-Ever Serverless Database Discounts

Serverless databases finally get commitment pricing:

• Aurora Serverless v2: up to 35% off
• DynamoDB on-demand: approximately 18% off

No more choosing between serverless flexibility and cost optimization.

3. Simplified Cost Management

Instead of managing separate Reserved Instances for each database service, you manage one savings plan.

One commitment. Nine services. Automatic application.

4. No Upfront Payment Required

The one-year term requires no upfront payment. You pay the discounted hourly rate as you go.

This reduces cash flow impact compared to partial or all-upfront Reserved Instances.

5. Protection During Architecture Changes

Modernizing your database layer? Database Savings Plans protect your savings during the transition.

• Migrating RDS to Aurora: covered
• Moving provisioned to serverless: covered
• Changing database engines: covered
• Consolidating databases: covered

4 Limitations You Need to Know

Database Savings Plans aren't perfect. Here's what they don't cover:

1. Older Instance Families Are Excluded

Not eligible: m5, r5, t3, t4g instance families

If you're running databases on these older instance types, you still need Reserved Instances for discounts.

What to do: Plan migrations to newer instance families (r6g, r7g, x2g) to unlock Savings Plan eligibility.

2. One-Year Term Only

Unlike Compute Savings Plans (which offer 1-year or 3-year terms), Database Savings Plans are one-year only.

No 3-year option means you can't access deeper long-term discounts.

3. ElastiCache Redis and Memcached Excluded

Only ElastiCache for Valkey qualifies. Redis and Memcached workloads need separate Reserved Instances.

4. No Convertible Option

Reserved Instances offer convertible options that let you change instance attributes. Database Savings Plans don't need this because flexibility is built in, but there's no equivalent "convertible" tier.

How to Calculate Your Optimal Commitment

Getting your commitment level right is critical. Too high wastes money. Too low leaves savings on the table.

Step 1: Analyze Current Database Spend

Pull your last 3-6 months of database costs from AWS Cost Explorer. Look for:

• Total eligible database spend (Aurora, RDS, DynamoDB, etc.)
• Hourly usage patterns (peak vs. baseline)
• Percentage on eligible vs. ineligible instance families

Step 2: Identify Your Baseline

Find your minimum consistent hourly spend across eligible services. This is your safe commitment floor.

Example:

• Peak hours: $15/hour database spend
• Off-peak hours: $8/hour database spend
• Safe baseline commitment: $8/hour

Step 3: Apply the 80% Rule

A conservative approach: commit to 80% of your baseline.

Using the example above: $8 x 0.80 = $6.40/hour commitment

This provides a buffer for usage variability while still capturing most available savings.

Step 4: Calculate Expected Savings

Multiply your commitment by the average discount rate:

$6.40/hour x 8,760 hours/year x 20% discount = $11,212 annual savings

Database Savings Plans vs. Reserved Instances

When should you use each? Here's a decision framework:

The Hybrid Approach

Most organizations should use both:

1. Database Savings Plans for modern, flexible, and serverless workloads
2. Reserved Instances for legacy instance types and maximum discounts on stable workloads

This combination maximizes savings while maintaining flexibility.

5 Best Practices for Implementation

1. Start Conservative, Scale Up

Don't commit 100% of your eligible spend on day one.

Start at 50-60% of your baseline. Monitor for 2-3 months. Add commitment as you confirm usage patterns.

2. Use a Laddering Strategy

Stagger your commitments with different start dates:

• Month 1: Commit 40% of baseline
• Month 4: Add another 20%
• Month 8: Add final 20%

This creates rolling expiration dates and maintains flexibility.

3. Separate Stable vs. Variable Workloads

Identify which databases have predictable usage (good for Savings Plans) versus highly variable usage (keep on-demand).

4. Review Quarterly

Database usage patterns change. Review your Savings Plan utilization quarterly and adjust future commitments accordingly.

5. Factor in Migration Plans

If you're planning to migrate off certain databases, exclude that spend from your commitment calculation.

Common Mistakes to Avoid

Over-committing: Committing more than your baseline hourly usage wastes money. Unused commitment still costs you.

Ignoring instance families: Assuming all databases qualify. Check that your instance families are eligible before committing.

Forgetting excluded services: ElastiCache Redis/Memcached need separate Reserved Instances.

Not monitoring utilization: AWS provides Savings Plans utilization reports. Check them monthly.

One-size-fits-all approach: Different workloads need different strategies. Analyze each database's usage pattern.

Summary: Is a Database Savings Plan Right for You?

Choose Database Savings Plans if:

• You run Aurora, RDS, or DynamoDB on modern instance families
• You use Aurora Serverless v2 or DynamoDB on-demand
• Your database architecture may change over the next year
• You want simpler cost management across multiple database services

Stick with Reserved Instances if:

• You run databases on m5, r5, t3, or t4g instance families
• You use ElastiCache Redis or Memcached
• You want maximum discounts with 3-year commitments
• Your workloads are stable and predictable

Use both if:

• You have a mix of modern and legacy database infrastructure
• You want to maximize savings across your entire database estate

The flexibility of Database Savings Plans makes them valuable for organizations embracing serverless, modernizing infrastructure, or operating across multiple regions. Combined with Reserved Instances for legacy workloads, they form a complete database cost optimization strategy.

FAQ Section

What are AWS Database Savings Plans?

AWS Database Savings Plans are commitment-based discounts for AWS managed database services. You commit to a consistent dollar-per-hour spend for one year, and AWS provides discounted rates of up to 35%. Unlike Reserved Instances, the discount applies flexibly across instance sizes, families, database engines, and regions.

How much can I save with AWS Database Savings Plans?

Savings range from 12% to 35% depending on the database type. Aurora Serverless v2 (IO-Optimized) offers the highest discount at up to 35%. Instance-based RDS and Aurora on modern families save around 20%. DynamoDB on-demand saves approximately 18%, and DynamoDB provisioned saves about 12%.

What AWS database services are covered by Database Savings Plans?

Nine services are covered: Amazon Aurora (including Serverless v2), Amazon RDS (all engines), Amazon DynamoDB (on-demand and provisioned), Amazon ElastiCache for Valkey, Amazon DocumentDB, Amazon Neptune, Amazon Keyspaces, Amazon Timestream, and AWS Database Migration Service.

Are serverless databases eligible for Database Savings Plans?

Yes. Database Savings Plans are the first commitment discount option for serverless databases. Aurora Serverless v2 can receive up to 35% discounts, and DynamoDB on-demand mode qualifies for approximately 18% savings.

Should I use Database Savings Plans or Reserved Instances?

Use Database Savings Plans for modern instance families, serverless workloads, and when you need flexibility across regions or engines. Use Reserved Instances for older instance families (m5, r5, t3, t4g), ElastiCache Redis/Memcached, or when you want maximum 3-year discounts on stable workloads. Most organizations benefit from using both.

What instance families are NOT eligible for Database Savings Plans?

Older instance families including m5, r5, t3, and t4g are not eligible. Databases on these instance types require Reserved Instances for commitment-based discounts.

Is there a 3-year option for Database Savings Plans?

No. Currently, Database Savings Plans only offer a one-year term. If you want deeper discounts through a 3-year commitment, you'll need to use Reserved Instances for eligible workloads.

Do I need to pay upfront for Database Savings Plans?

No upfront payment is required. You pay the discounted hourly rate as you use eligible database services throughout the one-year term.