Rightsize EC2 and RDS Instances
Overprovisioned instances is the biggest cause of cloud waste so should be a priority on any AWS cost efficiency checklist. AWS provides a number of tools to rightsize EC2 and RDS instances and even has a “rightsizing recommendations” tool. However, these use utilization metrics over an eight day period to make their recommendations, so may not be suitable in all use cases - and they won´t recommend when a general use instance should be rightsized to a memory or compute optimized instance.
Upgrade/Downgrade Instances to the Latest Generation
Every now and then, AWS releases upgraded instances with improved performance and functionality. Upgrading existing instances to “same capacity” new generation instances won´t necessarily increase cost efficiency, but there will likely be opportunities to downgrade existing instances to “smaller capacity” new generation instances to benefit from the improved performance and functionality at less cost. In order to identify candidates for downgrading, look for utilization of 40 percent or less.
Delete Unattached EBS Volumes
When you launch an EC2 instance, an Elastic Block Storage (EBS) volume is automatically. If you don´t check the box to automatically delete the EBS volume when the EC2 instance is terminated, AWS will continue to charge for it even though it is not being used. If you have EBS volumes marked as “available” in the AWS console, you are likely paying for storage you no longer need. If you´re not sure whether or not you need the data stored on the EBS volume, take a snapshot before deleting the resource.
Delete Obsolete Snapshots
Individually, EBS snapshots do not cost very much; but, if you continue paying for snapshots when they are no longer required, it´s not a very efficient use of AWS costs. Most of the time, snapshots are only used for recovery purposes when something goes wrong, and then only the most recent snapshots will be required. Therefore, check through your inventory for snapshots that are obsolete and delete those you no longer need.
Delete Unattached Elastic IP Addresses
Elastic IP addresses have a unique pricing structure inasmuch as when the instances to which they are attached to are running, they are free of charge. However, in order to encourage businesses to release IP addresses once they have finished using them, AWS charges a fee for IP addresses not released back into the pool. The fees add up quickly, and although Elastic IP addresses are difficult to find in the AWS System Manager or AWS Console, the effort put into looking for them can increase cost efficiency.
Terminate Zombie Assets
The term “zombie assets” can relate to any unused assets that create cost inefficiencies; but, in the context of our AWS cost efficiency checklist for 2020, it relates to EC2 and RDS instances that have been abandoned or were created in a test environment and then forgotten about. It is also worth looking out for unused and underused Elastic Load Balancers in your production environment that could be good candidates for termination.
Apply Start/Stop Schedules to Non-Production Resources
Whilst on the subject of test environment, this is an appropriate point in our AWS cost efficiency checklist to remind businesses you don´t pay for what you use in the cloud, you pay for what you provision. Consequently, if you provision instances for a test environment, you will be charged for them all the time they are left running. Typically business can save 65 percent by applying a start/stop schedule to non-production resources, or more with a utilization-based start/stop schedule.
Make Sure RIs and Savings Plans are being Fully Utilized
Many industry observers write about how businesses can save money by purchasing Reserved Instances (RIs) and Savings Plans, but few mention they can also cost businesses money if not utilized beyond a break-even point. The management of AWS´ discount programs has become more complicated over the years due to the introduction of Convertible Reserved Instances, EC2 Instance Savings Plans, and Compute Savings Plans, but getting on top of RI/Savings Plan management and ensuring the discount programs are fully utilized at all times is the most efficient use of cloud spend.
Purchase Reserved Nodes for Redshift and ElastiCache Services
EC2 and RDS instances are not the only resources that offer the opportunity to reduce costs by purchasing reservations. If your businesses uses AWS Redshift and ElastiCache, you can also purchase Reserved Nodes for these services in order to make more efficient use of cloud spend. Reserved Nodes work in much the same way as Reserved Instances inasmuch as you can purchase them over 1-year or 3-year terms paying all cost upfront, partially upfront, or no-upfront.
Move Infrequently-Accessed Data to Lower Cost Storage Tiers
AWS offers five tiers of S3 storage at different price points depending on propensity to risk and the frequency with which data is accessed, plus a further Intelligent Tiering service which automates where data is stored based on how often it is accessed during the previous thirty days. Although Intelligent Tiering can ensure businesses are not paying too much for some storage costs, manual operations are still required to ensure data is stored in compliance with data protection regulations.
How CloudHealth can help tick items off the AWS cost efficiency checklist
CloudHealth is an advanced cloud management platform that not only provides total visibility of the AWS Cloud so businesses can identify cost inefficiencies, but also gives businesses the opportunity to automate many optimization processes. By taking advantage of CloudHealth´s automation capabilities, businesses can eliminate concerns about unused or underutilized resources, Reserved Instance management, and data being placed in a storage tier where it is not instantly accessible.