We work with companies who are building software for all kinds of industries and purposes. However, there are some patterns we see across all of them when it comes to their cloud bills and what they might be overspending on. Here are five of the most common AWS cost optimization opportunities we see recurring within organizations, along with some pro tips on how engineers or other team members can fix them.
#1: Misconfigured NAT Gateways
If data is repeatedly or unnecessarily transitioning between AWS regions, NAT Gateway charges can get costly. Typically, if your NAT Gateway data transfer costs exceed 10% of your cloud bill, you should ask whether these costs are reasonable (they may well be if you’re operating a high-traffic website, but not if you’re running an average application). AWS provides some guidance on NATGateway Optimization, which can be found here.
#2: Overuse of AWS Snapshots
Before AWS introduced centralized, automated backup and restore capabilities, AWS snapshots were commonly used for data backup and archiving. In some legacy architectures, engineering teams may unintentionally be spending up to 20% of their cloud bill on snapshots! Some engineers tend to err on the side of caution and back everything up, but not everything requires a snapshot (nor does that snapshot need to be stored forever.)
If your organization is still using snapshots, it may be smart to investigate where they’re needed and why. Some information may be necessary to keep for a limited amount of time for compliance purposes, for example. However, you may be storing old snapshots for absolutely no reason, or snapshotting unnecessary information. After you’ve done an initial audit, you may want to investigate more modern backup capabilities, such as those built into AWS now, and ditch the snapshots once and for all.
#3: Outdated Compute
As AWS introduces new product families and deprecates others, many organizations may be left with the technical debt of outdated compute. While it may not always seem like the most pressing concern to upgrade, new product families tend to perform better and cost less than legacy alternatives.
As we noted in a previous blog post about proactive architectural changes, upgrading compute may or may not be a quick fix. There’s usually testing and script updates required to move to a new family, but it’s a key opportunity to improve performance and save money in the cloud. If you set up cloud cost alerting, you may not want to get an alert on every instance of outdated compute, but we recommend keeping an eye on the ratio and noting anytime it goes above 10% for a given account, as this is likely a valuable opportunity to upgrade.
#4: Too Much Overhead on AWS Management Services
Many services could fall into the category of AWS Management Services, including Amazon CloudWatch, AWS CloudTrail, AWS Config, AWS X-Ray, and more. Depending on the type of system, there may be more spending required in the area of AWS Management. For example, cloud-native or serverless infrastructures may have a higher percentage of overall spend dedicated to AWS Management (e.g. higher than 10 percent of the overall cloud bill), and that’s okay.
For most organizations, it makes sense to see if the percentage of overall spend on overhead is necessary. In some cases, investing in certain AWS services could prevent unnecessary expenses. For example, if CloudTrail events are missing, your organization may not be properly auditing API events across your infrastructure. Additionally, you may want to see if there are any redundancies across your third party management services. For example, are you paying for both CloudWatch and another log aggregation tool? Regardless, it’s worth it to periodically check that your spending on AWS Management Services is strategic.
#5: Tax Setup
As the old saying goes, the only things certain in life are death and taxes. However, your organization may be overpaying for unnecessary AWS taxes if your account is set up in a certain location. While this topic is more the realm of accounting and legal, it’s worth noting that certain settings can be changed within AWS to minimize the tax impact for your organization.
Usually, when a new account is set up, the account owner enters a billing address that determines how the organization is taxed. For businesses with multiple addresses or locations, it may make sense to configure the AWS account for the state (if in the US) or region with the lowest tax impact.
AWS Cost Optimization: Catch Cost Mistakes Fast
The best defense against accidental AWS overspending is full visibility into live cloud costs and real-time cost anomaly detection. With access to AWS cost optimization tools, and the right data at the right time, teams can improve cost optimization, while increasing efficiency and application performance at the same time. The more proactive you can be about cloud cost optimization management, the less of a shock you (and your finance team) will get when you receive the monthly AWS bill.
CloudZero is the first AWS cost optimization tool to put cloud cost data in the hands of engineers in a format that is useful and timely for their efforts. To learn more about CloudZero’s AWS cost optimization capabilities, get started here.
Learn More About CloudZero
CloudZero is the first real-time cloud cost platform designed specifically for engineering and DevOps teams.