AWS RDS: Quick Tips for Cutting Costs

What is Amazon RDS?

Amazon Relational Database Service (Amazon RDS) makes it easy to set up and manage relational databases in the cloud. It offers cost-efficient and scalable options, handling tedious database administration tasks such as scaling computing resources, storage for a DB instance, and managing automatic backups and software updates.

Amazon RDS supports six widely used database engines—Amazon Aurora (with MySQL or PostgreSQL compatibility), MySQL, MariaDB, PostgreSQL, Oracle, and SQL Server—and allows for deployment on-premises via Amazon RDS on AWS Outposts.

While Amazon RDS streamlines database management, costs can still escalate. This blog will explore the advantages of using Amazon RDS, identify the main factors that drive costs, and discuss how solutions like Cloudchipr can assist in managing RDS expenses effectively.

Amazon RDS cost factors

Amazon RDS pricing includes multiple components:

1. DB Instance: Charges are based on the compute capacity used per hour for your DB instance.
2. Database Storage: Costs are charged for storage capacity on a per GB-month basis.
3. I/O: Payment is required for IOPS and throughput that exceed the baseline capability, priced per IOPS-month and MB/s-month.
4. Backups: Additional backup storage that exceeds 100% of your total database storage for a region incurs charges per GiB-month.
5. Data Transfer: While incoming data transfers to RDS from the internet are free, outgoing data transfers can be costly, charged per GB.

To optimize costs for Amazon Relational Database Service (RDS) on AWS, consider implementing the following strategies:

Switch to Amazon RDS Graviton-based instances:

Shift from Intel to Graviton processors for your RDS instances. This switch can reduce costs by nearly 10% while maintaining the same resources and requiring no modifications to your application. Graviton instances, which are ARM-based processors, can offer up to a 52% improvement in price/performance for RDS open-source databases, depending on the database engine, version, and workload type. For detailed guidance on making this change, refer to this AWS blog. Note that instances ending with "g" indicate they are Graviton-based.

Use Amazon RDS Reserved Instances

Reserved Instances let you book a database instance for anywhere from one to three years, which can greatly reduce your costs.

When you choose a Reserved Instance, you can save a lot on your expenses. For instance, booking for one year can cut your costs by up to 45%.

Amazon RDS Reserved Instances are available in three payment options:

• All Upfront: This option lets you pay all of the Reserved DB Instance fees upfront, and as a result provides the largest discount compared to on-demand pricing.
• Partial Upfront: This option lets you pay a portion of the upfront payment. You will then be charged the discounted rate for the duration of the Reserved Instance term.
• No Upfront: With this option you don’t have to pay any fees upfront, but pay for the discounted hourly rate for the duration of the term. This option provides the smallest discounts and is only available for one year term.

If your system usage is consistent and your setup is stable, using Reserved Instances can help you save a lot of money. Also, if you have Reserved Instances that you're not using, you can sell them on the Amazon Marketplace.

Selecting the right storage option

Amazon RDS offers three storage options:

1. General Purpose SSD: This is a cost-effective storage option suitable for a variety of workloads, particularly effective for medium-sized DB instances. The performance baseline offers 3 IOPS per GiB, starting at a minimum of 100 IOPS and capping at 16,000 IOPS, with the ability to burst beyond the baseline. Storage capacity options range from 20 GiB to 64 TiB. Billing is based solely on storage usage for the gp2 type, whereas the gp3 type also includes charges for IOPS and throughput. This is the most budget-friendly option.
2. Provisioned IOPS SSD: Ideal for I/O-intensive applications that demand low latency and consistent throughput. Storage can be scaled from 100 GiB to 64 TiB, with IOPS ranging from 1,000 to 256,000. Costs are incurred for both the IOPS and the storage capacity you provision, making this the pricier choice.
3. Magnetic Storage: Maintained primarily for legacy support, this option is generally not recommended for new projects due to its lower storage capabilities compared to other types. It serves those with specific backward compatibility requirements.

It's important to select a storage type that matches your workload needs. If the IOPS rate is too low, your database may respond slowly. On the other hand, choosing a storage option with too high of an IOPS rate can lead to unnecessary costs.

You can track your storage usage with AWS CloudWatch metrics and set up alerts for when usage reaches certain levels. AWS offers this monitoring service at no extra charge, so it’s beneficial to regularly check metrics like memory, CPU, network I/O, and open sessions to optimize your storage settings.

Use read replicas and simplify queries

Using read replicas can make reading data from your databases faster and reduce stress on your main database.

For example, imagine your database has a very large column and you need to retrieve all the data in it with a SELECT query. Doing this in a busy production environment can slow down or even stall your database. However, if you use read replicas, the query will pull data from the replicas instead. This way, your main database isn’t overloaded by the query, keeping it running smoothly.

Set up archival and purging strategies:

Amazon RDS supports cost optimization through effective data archival and purging techniques. Here are key practices to consider:

a. Data Archival: Identify data that is older or accessed less frequently for archival to minimize your storage use. Use Amazon S3 or Glacier for long-term storage of this data. Archiving data to these services reduces your RDS storage costs while keeping the data accessible when needed. For guidance on exporting data from RDS to S3, refer to this AWS documentation.

b. Data Purging: Conduct regular reviews of your RDS databases and remove any data that is no longer necessary. Purging redundant or outdated data not only cuts down on storage expenses but also enhances database performance and the efficiency of queries.

Manage Amazon RDS snapshots and backup costs

When using Amazon RDS, the size of your database snapshots is determined by the actual data size in your database, not the total allocated storage size. This is important to note for optimizing your backup storage and understanding the associated costs.

Snapshot Details:

• The first snapshot taken is a full backup of your database. For instance, if your database's total allocated storage size is 20 GB but only 10 GB is in use, the first snapshot will only be 10 GB.
• Subsequent snapshots are incremental, meaning they only capture what has changed since the last backup. Depending on the modifications made, these snapshots could range from 1 GB to 3 GB.

Backup Storage Costs:

• Amazon RDS does not charge for backup storage up to 100% of your total database storage for the region. In this example, you are entitled to 20 GiB of backup storage per month at no extra cost.
• Any backup storage that exceeds this limit is charged at a rate of $0.095 per GiB per month. The size of these backups will fluctuate based on your storage usage and the rate of change in your data.

Review your backup and retention settings for RDS instances. Tailor the backup retention period and frequency to align with your business needs. By shortening the backup retention period or reducing the frequency, you can lower storage costs while still safeguarding your data. To adjust these settings, modify the instance and navigate to the “Additional Configuration” section.

Simplify Amazon RDS cost management with Cloudchipr

Balancing cost efficiency with performance in Amazon RDS can be challenging. Often, you may find yourself overspending due to excess resources like compute, memory, and storage, and resources that aren't being used, such as backups and snapshots. These require careful handling or removal to cut unnecessary costs.

Cloudchipr makes managing these expenses easier. Connecting directly with AWS helps you monitor your spending effectively. Cloudchipr combines all cloud-related costs, even from multiple cloud environments, into one simplified platform.

"Live Usage & Mgmt" feature:

Cloudchipr’s “Live Usage & Management” tab enables you to monitor your databases effectively. Enhanced filtering capabilities allow you to quickly identify and address underutilized or overlooked databases, easily assessing their financial impact.

In the image below, you can see an example of filters designed to display databases whose CPU utilization has been below 10% for the past seven days and have had fewer than 300 connections.

Cloudchipr also provides insights into RDS snapshots. Enhanced filtering lets you easily identify unnecessary snapshots and determine how much they add to your bill.

Here is an example of filtering snapshots that were created manually.

"Off-Hours" feature:

Part of Cloudchipr's "Automations" feature, the Off-Hours capability allows you to schedule DB instances in the development environment to stop/start at your preferred time. This helps optimize cloud operations by reducing costs incurred from running resources when they're not needed.

"Workflows" feature:

Another aspect of the "Automations" feature is the ability to delete temporary or sandbox environment databases. This helps in removing unused instances, thereby further optimizing your cloud operations by eliminating unnecessary data storage and associated costs.

In summary, leveraging Amazon RDS can significantly streamline database management while ensuring scalability and cost-efficiency. This blog has explored various strategies to optimize RDS costs effectively, such as transitioning to Graviton-based instances, utilizing Reserved Instances, selecting the appropriate storage options, implementing read replicas, and applying archival and purging strategies. Additionally, managing snapshots and backup costs is crucial in maintaining a cost-effective RDS deployment.

Cloudchipr enhances these efforts by providing a comprehensive platform that simplifies the management of RDS resources. Its features, like the “Live Usage & Management” and “Automations”, enable proactive monitoring and management of databases and snapshots, ensuring resources are optimized, and costs are minimized. By integrating these strategies and tools, businesses can achieve a balanced approach to cost management and performance optimization in their RDS environments.