The content in this blog is outdated and we cannot reliably say it is still accurate with the speed in which the cloud industry moves. But don’t worry—below are more recent, up-to-date blogs.
What does ‘rightsizing’ mean?
The easiest way to understand rightsizing is to think of it as modifying your cloud infrastructure to equate it with actual demand.
While many different cloud services can be ‘rightsized’, Relational Database Service (RDS) or Elastic Compute Cloud (EC2) are most commonly referenced. This is due to the fact that monthly cloud costs are most often driven by these services.
Why should I be rightsizing?
Rightsizing helps you identify underutilized assets, allowing you to make an informed decision to adjust the assets assigned to the instance, a.k.a, ‘rightsize’ it, or decommission it.
So, why are you stuck paying for this?
AWS is similar to a utility provider, not unlike your local electric company. For example, if you forget to turn your lights off when nobody is home, you are still responsible for paying for the usage. AWS is similar, in that if you deploy compute and database instances that are not efficiently utilized, you still pay...despite how few people use them.
Which metrics should I pay attention to, in order to rightsize?
This is a common question, and I always answer with ‘it depends’. You might have Key Performance Indicators (KPIs) in your business that can be leveraged to determine what resources are provisioned.
Unsure where you should start? Leverage CPU, and Memory Utilization as a jumping off point. The good news is that you are able to get CPU utilization from CloudWatch (which is a free service from AWS). The bad news is that retrieving memory utilization data is a bit more involved. CloudWatch depends on the hypervisor for information, which means it isn’t able to obtain operating system metrics such as memory utilization.
So, what can you do? While there are a number of solutions, here are two of my favorites.
- Extend CloudWatch with custom metrics. AWS has a great write up to get you started.
- Leverage third-party SaaS tools, such as Datadog and CloudHealth, to rightsize your cloud with real-time performance data and analytics. Together, Datadog and CloudHealth bring you end-to-end visibility across every dimension of your cloud environment. The granular analysis and performance data enables you to make sound decisions when it comes to provisioning and rightsizing in the cloud.
Selecting the Right Instance Type
Amazon currently offers six instance types: Burstable, General Purpose, Compute, Memory (X and R), GPU, and Storage Optimized (Dense and I/O). It’s important to understand how your applications use resources (CPU and Memory) prior to moving to AWS. Selecting the correct instance type for your workloads helps ensure that you’re getting the most performance for your application running on the AWS platform.
Cloud workloads are dynamic by nature, and the ability to scale infrastructure up and down on-demand is one of the most significant benefits of the public cloud. Cloud workloads are also self-service, enabling users who need to provision assets to easily sign in and spin up infrastructure. However, these benefits can come with challenges. Organizations may find that their compute instances are highly underutilized for two primary reasons: 1) Either a workload is no longer as resource-intensive as it was initially, or 2) the instance has been accidentally, or purposefully, over-provisioned. This can quickly drive up spend if it goes unchecked. I recommend continuously evaluating utilization and efficiency across your cloud environment and leveraging the resources available to make data-driven decisions.
About Brett Gillett
Brett is a results oriented technology enthusiast with a proven track record of designing and deploying large complex systems management and security solutions in a wide variety of corporate environments. As the AWS Practice lead at Softchoice and an AWS certified expert, Brett helps clients leverage Amazon Web Services (AWS) for their computing requirements. Brett enables clients to take advantage of cloud services such as AWS so they are able to concentrate on what differentiates their business from their competitors, rather than spend time on infrastructure management.