I wanted to be mindful of instances that were potentially overprovisioned or poorly matched to workload needs. Below were my steps:
- Amazon Cloudwatch: I went to Cloudwatch (logo is in the banner of this article). I then viewed the service dashboard for EC2 (Image 1). I am currently only running one EC2 instance, otherwise I would’ve had the ability to select individual EC2 instances and also toggle for a specific timeframe.
- Amazon EC2 Resource Optimization Recommendations: This time I went into Cost Explorer and viewed recommendations. Optimization opportunities, estimated monthly savings, and estimated savings as a percent (%) were displayed.
- Downloading Amazon EC2 Resource Optimization CSV File and sort it to find quick wins: After downloading the .csv file, you can organize your table by identifying instances that are too small, recommended Instance Type 1 estimated savings, recommended action (ie terminate), and more.
- A Note on Amazon EC2 Rightsizing Best Practices: It is noted that best practices for EC2 Rightsizing include starting simple, rightsizing prior to a migration, “measure twice, cut once,” address rightsizing regularly and as a part of your systemic operations, test once and perform multiple rightsizings, combine reserve instances or savings plans with rightsizing to optimize savings, and “ignore burstable instance families (T types).”
Credit: This exercise was inspired & guided by a lab at wellarchitectedlabs.com