7 Ways to Prevent AWS Bill Aftershock

Most organizations have realized that they spend too much on acquiring and maintaining their on-premises data center facilities—particularly with regards to capital expenses for additional capacity—to support times when demand is at its peak. As such, there’s an ongoing shift to public cloud providers such as AWS that allows companies to reduce overall costs while increasing utilization rates of workloads.

Shutting down unused instances, especially during weekends and at the end of the work day, is the first step in optimizing AWS costs. Services such as Elastic Beanstalk and AWS OpWorks enable developers to rapidly and consistently deploy/redeploy applications without worrying about the configuration of the underlying infrastructure. Using AWS CloudFormation to define infrastructure as code makes it easy for developers to create templates of AWS resources for the rapid building/rebuilding of environments. Such an approach to cloud computing enables the concept of ephemeral workloads that makes it possible to delete and shut down unused AWS resources.

2. Track, monitor and analyze service usage.

CloudWatch and Trusted Advisor are built-in management and monitoring tools used to access instance metrics. The analysis of the collected data will enable you to review your workloads properly, turn off non-production instances and determine if your instance size needs to be scaled up or down. Trusted Advisor is an innovative tool that runs configuration checks to identify idle resources while providing real-time guidance to enable the provisioning of resources based on AWS best practices. It provides weekly updates that pinpoint areas where you need to take action, thus helping reduce overall costs, enhance performance, and increase security. The Amazon CloudWatch service can be used to set alarms, collate, track and monitor key metrics and log files, and react automatically to changes in AWS resource use that may indicate wastage. Through a simple API that sends and stores metrics critical to your application’s performance, CloudWatch allows you to also monitor custom metrics generated by your applications. This enables you to gain system-wide visibility into operational health, application performance, and resource utilization, allowing you to take rapid action to manage wasteful or inefficient workloads.

3. Select the right instance type.

Since individual instance families usually have different costs, reducing your AWS bill will require the use of the most cost-effective instance that is suitable for your application workload. To maximize workloads and minimize spend, consider the specific use case to determine the amount of memory and type of processing unit required. Only then can you choose an instance resource that is both suited to the use case and cost-effective. Also, you should periodically review your choice of instances to determine if they are ideal for your current workload.

With AWS, you can save money by adapting services to meet current business needs. There is no need to compromise or overprovision capacity since you can right-size services to meet your exact capacity requirements. AWS allows you to select the services that meet existing business needs. And when business needs change, you can easily adapt service options to meet new requirements without fear of incidental fees or penalties. You can also maintain optimal performance while reducing costs by running multiple service options concurrently.

5. Leverage spot instances.

Using spot instances is another way to save money on non-stateful workloads. Spot instances are ideal for situations where you need large amounts of compute capacity to handle workloads but are unconcerned about possible interruptions. You should, however, have a mechanism in place for dealing with interruptions that may arise. To acquire spot instances, you simply need to bid on spare Amazon EC2 computing capacity. Such capacity is usually available at reduced prices or comes with discounts that are much lower than on-demand pricing. With spot instances, you can reduce the overall cost of running your applications or increase your application’s throughput and compute capacity on the same budget.

6. Purchase reserved instances.

For services like Amazon EC2 and Amazon RDS, you can reduce your AWS bills by using a Reserved Instance instead of its on-demand equivalent. A reserved instance gets you a significantly lower hourly rate in exchange for a one or three-year commitment to purchase a reserved capacity. This enables you to save over 70% of the price of an equivalent on-demand capacity. If you are unable to use up the purchased capacity, you can sell it on the reserved instance marketplace or buy one with a shorter duration. Reserved instances are available in three payment options:

• NURI (no upfront payment)
• PURI (partial up-front payment) and
• AURI (all up-front payment)

When purchasing reserved instances, you get a greater discount when you make a greater upfront investment. This means that using the AURI option (all upfront payment) will net you the largest discount. Choosing reserved capacity helps your organization to comply with policies requiring longer-term commitments, minimize risks, and predictably manage its budgets.

7. Evaluate alternatives.

There are other ways to optimize your AWS billing that don’t involve the above methods. These include:

• Using the appropriate storage classes and autoscaling
• The use of Consolidated Billing

AWS can drastically reduce the cost of running your platform and infrastructure services, but only if you use it correctly and properly. Organizations must look beyond the pricing advantages of AWS over on-premises data centers and use the above methods to use its services effectively while lowering their AWS bill.