Amazon RDS / Aurora Review 2026: AWS's fully managed relational database — from $0.016/instance-hour to per-second Aurora Serverless scaling
Affiliate disclosure: this review contains affiliate links — we may earn a commission if you sign up, at no cost to you. Ratings are our own editorial scores.
Amazon RDS / Aurora
Pros
- Entry db.t4g.micro is genuinely cheap at $0.016/hr (~$11.52/mo) and Graviton instances cut compute cost vs x86
- Aurora Serverless v2 bills per-second and scales from 0 ACUs, so idle dev/test databases cost almost nothing
- Fully managed: automated backups, patching, Multi-AZ failover, and up to 15 read replicas remove ops overhead
- Broad engine choice — MySQL, PostgreSQL, MariaDB, Oracle, SQL Server, plus Aurora's faster compatible engines
Cons
- Usage-based billing is hard to predict; storage, I/O, backups, and data-transfer stack on top of instance-hours
- Aurora Standard's $0.20 per million I/O requests can shock write-heavy workloads unless you switch to I/O-Optimized
- Data-transfer out at $0.09/GB and Multi-AZ roughly doubling compute inflate real-world monthly costs
- Free tier for accounts created after Jul 15 2025 is now credit-based ($100–$200), not 750 free hours, so it expires faster
Best for: Teams already on AWS wanting a managed relational database, Variable or spiky workloads that benefit from Aurora Serverless v2 scale-to-zero, Production apps needing Multi-AZ high availability and read replicas.
What is Amazon RDS / Aurora?
Amazon RDS is the managed relational database service at the center of the AWS data stack, and Aurora is its high-performance branch. AWS describes RDS as optimized for total cost of ownership, automating the undifferentiated tasks that consume database administrator time: provisioning, configuring, backing up, patching, failure detection and repair. You pick an engine — PostgreSQL, MySQL, MariaDB, SQL Server, Oracle or Db2 — choose an instance class and storage type, and a database is running in minutes.
Aurora shares that console and operational model but swaps in fault-tolerant distributed storage that keeps data durable across three Availability Zones while you pay for one copy. The vendor positions Aurora as fully MySQL- and PostgreSQL-compatible with six times the throughput of those community engines, and Aurora DSQL extends the family with a PostgreSQL-compatible distributed SQL option. Read Amazon RDS and Aurora as one product line with two performance tiers rather than competing services.
Engine choice and deployment options
Breadth is the point. Beyond the standard managed service, Amazon RDS offers RDS Custom when an application needs privileged access to the underlying operating system, and RDS on AWS Outposts when a managed instance must sit in your own facility. Migration is a supported path: AWS Database Migration Service and the AWS Schema Conversion Tool handle homogeneous and heterogeneous moves, while Babelfish for Aurora PostgreSQL runs legacy SQL Server applications with minimal code change.
Aurora Serverless and elastic capacity
Aurora Serverless is the piece worth studying if traffic is lumpy. Capacity is measured in Aurora Capacity Units billed per second, starting at zero ACU and adjusting in increments as small as half a unit, so staging clusters and seasonal workloads stop paying for idle provisioned hardware. AWS reports instant scaling to hundreds of thousands of transactions per second. Those Serverless clusters sit inside the wider Aurora feature set, which includes Multi-AZ deployments, up to 15 Aurora replicas for read scalability, and Aurora Global Database for cross-Region replication.
Operational work the service absorbs
This is where Amazon RDS repays its premium over a self-hosted instance. Automated backups provide point-in-time recovery, snapshots cover manual checkpoints, and Multi-AZ deployments run either one standby or two readable standbys. RDS Blue/Green Deployments make major-version upgrades far less frightening by staging a synchronized green environment before the cutover. CloudWatch Database Insights and Amazon DevOps Guru for RDS surface bottlenecks, while encryption at rest and in transit, Amazon VPC isolation, resource-level permissions and Amazon GuardDuty cover the security baseline auditors ask about.
Performance, analytics and AI workloads
Scaling levers are granular: push-button compute scaling, storage that grows on its own, Read Replicas, Amazon RDS Proxy for connection pooling, and Optimized Writes and Optimized Reads for write-heavy or memory-constrained workloads. Zero-ETL integrations with Amazon Redshift and Amazon SageMaker replicate transactional data for analytics without a pipeline to maintain. Both Aurora PostgreSQL and RDS for PostgreSQL store vectors and embeddings for similarity search, and the vendor reports 20x improved queries per second from Aurora Optimized Reads with pgvector_hnsw compared with pgvector_IVFFLAT.
Understanding the cost model
Pricing is metered rather than packaged. On-Demand Instances bill per second with no commitment, Reserved Instances trade a one- or three-year term for a substantial discount, and Database Savings Plans commit to an hourly spend instead. Aurora adds a configuration choice: Aurora Standard bills instances, storage and per-request I/O, while Aurora I/O-Optimized drops read and write I/O charges entirely and, the vendor says, saves money once I/O passes a quarter of your Aurora bill. Rates vary by Region.
Who should choose Amazon RDS / Aurora
Teams already building on AWS get the strongest case: the IAM, VPC, CloudWatch and Redshift integrations remove work that would otherwise be yours, and the Free Tier makes a first cluster cheap to evaluate. Enterprises escaping Oracle or SQL Server licensing have a genuine route out through Babelfish and AWS DMS. It is a poor fit for a solo developer who wants one flat monthly figure and no capacity planning, since instance hours, storage, I/O, backups and data transfer all meter separately.
Key features
| Feature | What it does |
|---|---|
| Managed engines | MySQL, PostgreSQL, MariaDB, Oracle, and SQL Server on RDS; Aurora adds MySQL/PostgreSQL-compatible engines. |
| Aurora Serverless v2 | Per-second ACU autoscaling that scales to zero when idle; $0.12/ACU-hr (Standard), 1 ACU ≈ 2 GiB RAM. |
| Automated operations | Backups, patching, Multi-AZ automatic failover, and up to 15 low-latency read replicas. |
| Flexible storage | gp2/gp3 general-purpose and provisioned-IOPS io2; Aurora storage auto-grows up to 128 TB. |
| I/O-Optimized pricing | Predictable option that removes per-request I/O charges for I/O-heavy Aurora workloads. |
| Free tier / credits | 750 instance-hrs/mo for 12 months (legacy accounts) or $100–$200 in credits for new signups. |
Amazon RDS / Aurora pricing
| Plan | Price | Included |
|---|---|---|
| Free Tier | $0 | 750 instance-hrs/mo of db.t3/t4g.micro + 20 GB SSD + 20 GB backups, 12 months (accounts before Jul 15 2025). Newer accounts instead get $100 + up to $100 more in credits. |
| RDS entry instance (db.t4g.micro)POPULAR | $0.016/hr | ~$11.52/mo compute for MySQL/PostgreSQL, us-east-1 (2 vCPU, 1 GB, Graviton2). Storage billed separately. |
| RDS storage (gp2/gp3 SSD) | $0.115/GB-month | General-purpose SSD; provisioned IOPS (io2) and backup storage beyond the free allotment cost extra. |
| Aurora Serverless v2POPULAR | $0.12/ACU-hour | Aurora Standard (MySQL/PostgreSQL); I/O-Optimized is $0.156/ACU-hr. Bills per-second, scales from 0 ACUs; 1 ACU ≈ 2 GiB RAM. |
| Aurora storage + I/O | $0.10/GB-mo + $0.20/1M I/Os | Aurora Standard. I/O-Optimized: $0.225/GB-month storage with $0 per-request I/O charges. |
How Amazon RDS / Aurora compares
| Alternative | How it differs |
|---|---|
| Google Cloud SQL / AlloyDB | GCP's managed MySQL/PostgreSQL; AlloyDB rivals Aurora on PostgreSQL performance and scaling. |
| Azure Database for PostgreSQL/MySQL | Microsoft's managed equivalent, strongest for Azure-centric stacks and enterprise agreements. |
| Supabase / Neon | Serverless Postgres with generous free tiers and scale-to-zero; popular with startups seeking flat pricing. |
Amazon RDS / Aurora ratings on other platforms
Independent user ratings from third-party review sites, linked here for transparency. These are not our editorial score, are captured on the date shown, and may have changed since.
Frequently asked questions
How much does Amazon RDS / Aurora cost?
Amazon RDS starts at $0.016 per instance-hour for a db.t4g.micro (about $11.52/month) plus $0.115/GB-month for gp2 SSD storage in us-east-1. Aurora Serverless v2 bills $0.12 per ACU-hour on Aurora Standard, scaling per second from 0 ACUs. Because pricing is usage-based, your total depends on instance size, storage, and I/O.
Is Amazon RDS / Aurora free?
There is a free tier, not a permanently free product. Accounts created before July 15, 2025 get 750 instance-hours/month of db.t3/t4g.micro, 20 GB SSD storage, and 20 GB backups free for 12 months. Newer accounts instead receive $100 in credits plus up to $100 more for activating services, valid for one year.
Amazon RDS vs Aurora — what is the price difference?
RDS runs standard MySQL/PostgreSQL on fixed instances from $0.016/hour with $0.115/GB-month storage and no per-I/O charge on gp3. Aurora is AWS's cloud-native engine that costs more per instance but offers Serverless v2 at $0.12/ACU-hour scaling to zero. Aurora Standard adds $0.20 per million I/Os; the I/O-Optimized tier removes that charge.
How is Aurora Serverless v2 priced?
Aurora Serverless v2 charges per Aurora Capacity Unit (ACU) per second, at $0.12/ACU-hour on Aurora Standard or $0.156/ACU-hour on I/O-Optimized in us-east-1. One ACU is roughly 2 GiB of RAM with matching CPU and networking. It now scales down to zero ACUs when idle, so you pay only for capacity actually consumed.
What are the hidden costs of Amazon RDS / Aurora?
Beyond instance-hours, watch storage ($0.115/GB-month RDS, $0.10 Aurora), backup storage past the free allotment, provisioned IOPS, and $0.09/GB data-transfer out. Aurora Standard's $0.20 per million I/O requests can dominate write-heavy bills; Multi-AZ roughly doubles compute. Burstable instances also incur CPU-credit charges of $0.075/vCPU-hour above baseline.
Verdict
Buy RDS/Aurora if you're building on AWS and want a managed relational database that scales — the $0.016/hr entry instance and Aurora Serverless v2's scale-to-zero make it viable from hobby projects to enterprise. Skip it if you want flat, predictable monthly pricing or aren't on AWS: usage-based billing plus I/O and data-transfer charges can quietly outrun a fixed-price competitor like Supabase or a self-hosted database.
Facts verified against: aws.amazon.com, aws.amazon.com, aws.amazon.com, instances.vantage.sh, aws.amazon.com, aws.amazon.com, aws.amazon.com, aws.amazon.com, b0.p.awsstatic.com, b0.p.awsstatic.com (as of August 2026).
More in Cloud, Hosting & Dev
See all →
AI code-review tool that automates pull-request analysis
AI-first code editor with agentic coding and tab completions
Frontend cloud for deploying and hosting web apps