Some Amazon Web Services cost optimization best practices are presumably recognizable to AWS users, but not all of them. As a result, we have produced a list of the 10 best methods for optimizing AWS costs, as well as a solution that ensures Amazon Web Services costs remain optimized.
It’s not uncommon to see stories suggesting that organizations are overpaying in the cloud, that a significant portion of money is being squandered on unneeded services, or that millions of businesses provision resources with more capacity than they require. Rightsizing, scheduling, and purchasing Reserved Instances/Savings Plans for predictable workloads are the most typical “solutions” to reported concerns.
These three “solutions” are perhaps the most well-known AWS cost optimization best practices among AWS users, however they are not always the “best” best practices. They don’t always save a fraction of the cost that is claimed, while there are plenty of other, frequently missed, AWS cost optimization best practices that can save a lot more. This is a problem that we intend to address further below.
The top ten AWS cost-cutting strategies
- EC2 Instance Rightsizing
Given that we’ve already discussed rightsizing, scheduling, and Reserved Instances/Savings Plans, let’s get started with these three AWS cost optimization best practices. The goal of rightsizing is to match the size of instances to their workloads. Unfortunately, it does not function quite like that due to the way instances double in capacity for each rise in size.
When you move up one size, you double the capacity; when you go down one size, you half the capacity. As a result, rightsizing is only a desirable best practice if peak utilization does not exceed 45 percent. It’s still worthwhile to look at usage numbers to see if there are any chances to migrate workloads to various families (other than “General Purpose”) that better suit their needs.
- Setting up on/off times
It is worthwhile to schedule on/off timings for non-production instances such as those used for developing, staging, testing, and QA, as doing so will save you around 65 percent of the cost of running these instances if you utilize a “on” schedule of 8 a.m. to 8 p.m. Monday through Friday. However, it is feasible to save even more—especially if development teams operate in erratic patterns or at erratic hours.
You can use more aggressive schedules by evaluating utilization metrics to see when the instances are most commonly utilized, or you can employ an always halted plan that can be interrupted when access to the instances is required. It’s worth noting that even when instances are scheduled to be turned off, you’re still charged for EBS volumes and other components associated with them.
- Savings Plans and Reserved Instances Investing
Buying Reserved Instances is a simple approach to cut AWS expenditures. It can also be an easy way to increase AWS costs if you don’t use the Reserved Instance as much as you expected, purchase the incorrect type of Reserved Instance, or purchase a “standard” Reserved Instance only to find AWS prices fall by more than the reservation “saves” over the term of your reservation.
Rather than suggesting that purchasing Reserved Instances is one of the best practices for AWS cost optimization, we’ll suggest that effective management of Reserved Instances is a best practice for AWS cost optimization—effective management entails weighing all the variables before making a purchase and then monitoring utilization throughout the reservation’s lifecycle.
- Remove any unattached EBS volumes.
To return to Elastic Block Storage (EBS), when you launch an EC2 instance, an EBS volume is associated to it to serve as the instance’s local block storage. When you end an EC2 instance, the EBS volume is erased only if you checked the “delete on termination” box when you launched the instance. If the box is not ticked, the EBS volume remains and contributes to the monthly AWS fee.
Thousands of unattached EBS volumes may exist in your AWS Cloud, depending on how long your company has been operating in the cloud and the number of instances launched without the remove box being ticked. Even if your company is new to the AWS Cloud, it is undoubtedly one of our AWS cost optimization best practices to consider.
- Remove out-of-date photos
Snapshots are an efficient way to back up data on an EBS volume to an S3 storage bucket because they only back up data that has changed since the last snapshot, preventing data duplication in the S3 bucket. As a result, each snapshot contains all of the information required to restore your data to a fresh EBS disk (from the time the snapshot was taken).
Typically, you’ll only need the most recent snapshot to restore data if something goes wrong (though it’s advisable to keep snapshots for a couple of weeks depending on the frequency with which they’re taken), and while snapshots don’t cost much individually, deleting those you no longer need could save you thousands of dollars.
- Unattached Elastic IP addresses should be released.
Elastic IP addresses are public IPv4 addresses from Amazon’s pool that are assigned to an instance so that it can be accessed via the Internet. Because Amazon does not have an infinite number of IP addresses, businesses are limited to a maximum of five Elastic IP addresses per account. They are, however, changeless when coupled to a functioning service.
Exceptions to the free rule arise when you remap an IP address more than 100 times per month, or when you keep unattached Elastic IP addresses after terminating the instances to which they were associated. The fee for unattached Elastic IP addresses is only $0.01 per hour, but if fifty AWS accounts each hold back two IP addresses, that equates to $8,760 in waste each year.
- Update instances to the most recent generation
Because of the breadth of Amazon Web Services’ products and services, there are frequent announcements regarding how products have been changed or capabilities added to support certain services. When it comes to AWS cost optimization best practices, the announcements linked to newest generation instances are the ones to keep an eye out for.
When Amazon Web Services introduces a new generation of instances, they often outperform their predecessors in terms of performance and functionality. This means that you can either upgrade current instances to the most recent generation or shrink existing instances with questionable utilization metrics to achieve the same level of performance at a lower cost.
- Purchase Redshift and ElastiCache Services reserved nodes
One recent AWS statement highlighted how the Amazon Redshift and ElastiCache discount programs had altered. Businesses could previously purchase “Heavy Utilization” discounts in advance, but these have recently been adjusted to (nearly) mimic Reserved Instance purchases for EC2 and RDS instances.
Reserved Nodes for Redshift, ElasticCache, Redis, and Memcached Services can be purchased for 1-year or 3-year durations, with the option of paying the full cost upfront, partially ahead, or monthly. One thing to keep in mind is that in order to use reservations on the ElastiCache Service, you must first upgrade your Nodes to the most recent version.
- Dispose of zombie assets
The phrase “zombie assets” is commonly used to characterize any underutilized asset that contributes to the expense of operating in the AWS Cloud—many examples of typical zombie assets have already been presented (unattached EBS volumes, obsolete snapshots, etc.). Other assets in this category include instance components that were active when an instance failed to run, as well as unused Elastic Load Balancers.
One issue that enterprises frequently face when attempting to follow AWS cost optimization best practices is the difficulty in locating idle assets. Unattached IP addresses, for example, are notoriously difficult to detect in AWS System Manager or AWS Console, and one of the reasons AWS promotes CloudHealth is that it offers enterprises with complete visibility of their cloud settings.
- Relocate infrequently accessed data to lower-cost tiers.
Amazon Web Services now offers six storage levels at various price points. The most appropriate storage tier for data will be determined by criteria such as how frequently data is accessed (since retrieval costs apply to lower tiers) and how soon a firm would need to retrieve data in the case of a disaster (as it may take hours to retrieve from a lower tier).
The cost savings of keeping infrequently accessed, non-critical data in a lower cost tier can be significant. Storing up to 50TB of data in a regular S3 storage bucket costs $0.023 per GB per month (US East Region), whereas S3 Glacier Deep Archive storage costs $0.00099 per GB per month. The six storage layers are as follows:
Standard S3
Intelligent Tiering (S3)
Access to S3 is infrequent.
Access to S3 is infrequent (Single Zone)
Glacier S3
Deep Archive Glacier S3
AWS cost reduction is a continuing endeavor.
Applying best practices for AWS cost optimization is a continual endeavor. Your AWS Cloud must be monitored at all times to determine when assets are underutilized (or not used at all) and when possibilities to save money exist by deleting/terminating/releasing zombie assets. It’s also critical to keep track of Reserved Instances to guarantee they’re completely utilized.
Establishing a Cloud Financial Management approach might also aid in the cost-cutting effort. Cloud Financial Management (CFM), also known as FinOps or Cloud Cost Management, is a role that aids in the alignment and creation of financial goals, the promotion of a cost-conscious culture, the implementation of financial guardrails, and the enhancement of business efficiencies.
