Say you are just starting out with AWS EC2–how do you know what instance type works best with your application? Selecting just the right size can be difficult even for those who know EC2 and their applications well. It involves understanding how much resources your app needs under normal workloads and what it will need during sudden spikes or deviations, then choosing an appropriate instance. After that, you would have to monitor and review your choice to determine if it needs to be further right-sized.
To simplify the process, AWS launched a new service: AWS Compute Optimizer. Like the Cost Explorer Rightsizing Recommendations, this service helps you determine the instance type you need. However, the recommendations are only a guide. AWS Compute Optimizer employs machine learning to study your applications’ actual resource usage over time, and uses the data to give clear recommendations as to what to right-size.
- Bring down compute costs by up to 25% — Recommends alternatives to replace over-utilized or underutilized resources, thereby lowering your operating expenses.
- Raise the efficiency of your applications – You are given up to 3 recommendations from the list of more than 140 AWS instance types, allowing you to right-size your instances quickly.
- Get Recommendations for Multiple Users — AWS Compute Optimizer is connected to AWS Organizations, which allows you to view recommendations for all AWS user accounts under your master AWS Organizations account.
AWS Compute Optimizer uses data from AWS CloudWatch metrics to come up with its analyses. Basic metrics include CPU utilization, disk IO, and network IO, but you may also include operating system level metrics by installing CloudWatch agent on your EC2 instance. AWS Compute Optimizer will automatically consider these metrics when generating recommendations for you.
AWS Compute Optimizer is also capable of presenting on-demand pricing, reserved pricing, reserved instance utilization, and reserved instance coverage, and their projected efficiency, as part of its recommendations. This gives you greater insight as to what instance provides the best price for the performance you need.
As a caveat, please be aware that it may be difficult to use Compute Optimizer to right-size older applications as they may not run properly when placed under different hardware architecture. They may have additional driver requirements or won’t work at all. Check with your app vendor first before running them on different configurations. Also, be sure to test the app under with its newly-recommended instance before applying it to production.
Talk to us at copebit if you would like to get started with Compute Optimizer. We can help you bring down those AWS costs. Chat with us here.
Marco is a managing partner at copebit. He got seven AWS certifications. He has spent three years in Australia and has worked with AWS and DevOps technologies for the last 6 years.