Rightsizing AWS RDS Instances: A Step-by-Step Guide

AWS RDS and Aurora instances are powerful tools for managing your databases in the cloud. However, choosing the right size for your instance can be a challenging task. On one hand, you want to ensure that your instance has enough resources to handle the traffic your application generates. On the other hand, you want to avoid overprovisioning and save on costs. In this guide, we'll cover the best practices for rightsizing your AWS RDS or Aurora instance, including how to identify when to rightsize, pick the initial size, understand the different instance families, and resize RDS instances with minimum downtime.

Understanding AWS RDS and Aurora Instances

Before we dive into rightsizing, let's first understand the basics of AWS RDS and Aurora instances. RDS stands for "Relational Database Service" and is a fully managed service that makes it easy to set up, operate, and scale a relational database in the cloud. Aurora is a MySQL and PostgreSQL-compatible relational database engine that is fully managed and designed to be more performant than RDS. Both RDS and Aurora instances come in different sizes, or "instance types", that are optimized for different workloads.

Understanding AWS RDS instance types

The different instance types available can be divided into several families, each with their own unique characteristics. The most commonly used instance families are:

• db.r6g: This family is designed for memory-intensive workloads and is optimized for high memory-to-vCPU ratios. It is a good choice for memory-intensive applications such as in-memory databases, high-performance computing, and data-intensive workloads. For example, the db.r6g.large instance type has 2 vCPUs and 16 GiB of memory.

• db.t4g: This family is designed for low-cost, general-purpose workloads. It is a good choice for development and test environments, small-scale production workloads, and applications that don't require a lot of CPU power. For example, the db.t4g.micro instance type has 1 vCPU and 1 GiB of memory.

• db.m6g: This family is designed for general-purpose workloads and is optimized for a balance of compute, memory, and network resources. It is a good choice for most production workloads. For example, the db.m6g.large instance type has 2 vCPUs and 8 GiB of memory.

Picking the Initial Size for RDS

When you're setting up a new RDS or Aurora instance, it's important to pick the right initial size to ensure that it can handle the expected workload. Picking the wrong initial size can result in poor performance or unnecessary costs.

There are several factors to consider when picking the initial size:

• Workload: The initial size should be able to handle the expected workload, including the number of read and write requests, the size of the data, and the complexity of the queries.

• Traffic: The initial size should be able to handle the expected traffic, including the number of concurrent connections and the amount of data transferred.

• Growth: The initial size should be able to handle expected growth in the workload and traffic for the very near future, so that you don't have to resize the instance in just a few days.

To pick the initial size, you can use the AWS RDS and Aurora instance types and families as a guide. As mentioned earlier, the general rule is that the larger the instance type, the more resources it has, and the more it can handle.

For example, if you're expecting a moderate amount of read and write requests, with a moderate amount of data and simple queries, you can start with a db.t4g.medium instance type. If you're expecting a high amount of read and write requests, with a large amount of data and complex queries, you can start with a db.m6g.large instance type.

It's also important to monitor the performance of the instance after it's launched and make adjustments as needed. AWS RDS and Aurora provide several performance metrics, such as CPU utilization, memory usage, and I/O operations, that can be used to evaluate the performance of the instance and identify when to resize.

Step-by-step guide:

1. Analyze the expected workload, traffic, and growth.

2. Choose an appropriate instance type and family based on the analysis.

3. Use the AWS Management Console or AWS CLI to launch the instance.

4. Monitor the performance of the instance using AWS RDS and Aurora performance metrics.

5. Identify when to resize the instance by analyzing the performance metrics and comparing them to the expected workload, traffic, and growth.

Identifying when to rightsize RDS

One of the key steps in cost optimization is identifying when to rightsize your RDS or Aurora instance. Rightsizing involves adjusting the resources of your instance to match the actual usage of your workload. This can include adjusting the CPU, memory, and storage of your instance to better match the needs of your application. Several methods can be used to identify when to rightsize your instance:

Method 1: CloudWatch Metrics

• Step 1: Go to the CloudWatch Console.

• Step 2: Select the desired time range for your metrics.

• Step 3: Locate your RDS or Aurora instance and select the "DBInstanceIdentifier"

• Step 4: Check the metrics for CPU, Memory and Storage usage.

• Step 5: Compare the usage with the current configuration of your instance and determine if there is an opportunity to rightsize.

Method 2: Performance Insights

• Step 1: Go to the RDS or Aurora Console.

• Step 2: Select your RDS or Aurora instance.

• Step 3: Click on the "Performance Insights" button.

• Step 4: Check the metrics for CPU, Memory and Storage usage.

• Step 5: Compare the usage with the current configuration of your instance and determine if there is an opportunity to rightsize.

Method 3: SQL Query

• Step 1: Connect to your RDS or Aurora instance using SQL client.

• Step 2: Run the following query: "SHOW STATUS WHERE variable_name = 'Threads_connected'"

• Step 3: Check the value of the "Threads_connected"

• Step 4: Compare the value with the current configuration of your instance and determine if there is an opportunity to rightsize.

By using one or a combination of these methods, you can gain a better understanding of the usage patterns of your RDS or Aurora instance and make informed decisions about when to rightsize.

How to Rightsize an RDS instance

There are several methods for rightsizing your RDS or Aurora instance, each with its own advantages and disadvantages.

Option 1: Scale Up

One method is to scale up the instance by modifying its instance type to a larger one. This can be done by modifying the instance settings in the AWS RDS or Aurora console. This method increases performance, but it also increases costs. Do this if your instance is near or at capacity.

Option 2: Scale Down

The opposite of scaling up is scaling down. It consists of modifying the instance type to a smaller one. This can be done by modifying the instance settings in the AWS RDS or Aurora console. This method will save you money, but it decreases performance. Do this if your instance is underutilized.

Option 3: Scale Out

Another method is to scale out the instance by adding read replicas. This can be done by creating a read replica of the instance in the AWS RDS or Aurora console. This method can improve the performance of read-heavy workloads, but it will not affect write performance. Do this if your workloads are read-heavy, or if you have a separate read-heavy workload from the same database (such as reporting).

Changing RDS Instance Size Without Downtime

When resizing an RDS or Aurora instance, it's important to minimize downtime to ensure that your application continues to function smoothly. One way to achieve this is by using the "Multi-AZ" feature, which allows you to create a secondary replica of your database in a different availability zone. This replica can be used to failover to in the event of an interruption to the primary database.

Another way to minimize downtime is by using the "Read Replica" feature, which allows you to create a read-only replica of your database. This replica can be used to offload read traffic from the primary database, which can be especially helpful when performing maintenance or scaling operations.

Step-by-Step Guide:

• Step 1: Create a read replica of your RDS or Aurora instance. This can be done through the AWS Management Console or using the AWS CLI.

• Step 2: Verify that the read replica is in a "Ready" state.

• Step 3: Promote the read replica to be a standalone database. This can also be done through the console or CLI.

• Step 4: Update the endpoint of your application to point to the new standalone database.

• Step 5: Test your application to ensure that it is still functional and all the data is up-to-date.

• Step 6: If the test is successful, terminate the old RDS or Aurora instance.

By following these steps, you can minimize downtime and ensure that your application continues to function smoothly during the resizing process.

Conclusion

In conclusion, rightsizing an AWS RDS or Aurora instance can save you money and improve performance. By understanding the different methods, instance types, and how to minimize downtime, you can make the process of rightsizing your RDS or Aurora instance a lot more smooth and effective.

Remember to constantly monitor your instances, and don't be afraid to make changes when necessary.

Thanks for reading!

If you're interested in building on AWS, check out my newsletter: Simple AWS.

It's free, runs every Monday, and tackles one use case at a time, with all the best practices you need.

If you want to know more about me, visit my website, www.guilleojeda.com