Top 5 AWS Cost Saving and Optimization Strategies

There are plenty of reasons to move to the cloud, but one of the key benefits from public cloud providers like AWS Cloud is pay-as-you-go pricing model with elastic scalability to grow or shrink resource usage based on your business needs and growth.

However, as you start using AWS Cloud, you will notice that the cost spikes and spend trends going upwards. So it’s important to have right tools like Botmetric for cloud cost monitoring and analytics, which can keep a track and control of your AWS cloud costs.

This blog post outlines the top 5 AWS cost saving and optimization strategies. Here are our recommendations listed below:

Use New Instance Families and Types

If you are using AWS cloud for the past few years, there are chances of running old generation instance families like m1, c1, t1 etc. If so you should move to m3, t2 and c3 family instances as it provides better CPU/Memory combination at lower price compared to the old generation instances. We have seen up to 10 to 20% AWS cost saving for customers who adopt new generation instance families and get better CPU/memory combination for the given price point.

You can refer to one of our recent posts on how to choose AWS instance family for your business workloads.

Shutdown DEV/QA Workloads on Weekends

One of the smart strategies for cost saving is to shutdown the DEV/QA workloads during weekends. You may also shutdown few dev instances during non-working hours as well. This would require your services to be bootstrapped to start/stop automatically when you start/stop your EC2 servers. Cloud automation tools like Botmetric support automaton jobs feature to shutdown and start EC2 servers based on instance tags on Friday night and bring them online on Monday early morning. It only requires few hours or days of effort but it can help you save anywhere 15% to 20% of the EC2 costs for DEV/QA instances.

Remove Unused Resources

In any typical business, different teams like product, QA, IT, automation and developers are using AWS cloud for developing, testing, deploying and managing applications. We have seen many unused resources like volumes, elastic IPs, old snapshots, unused data in S3, idle load balancers, heavily under utilised machines etc. Tools like AWS Trusted Advisor and Botmetric provide insights on unused resources so that you can clean them up for saving cost. Botmetric also provides automation jobs to ensure you keep track of cloud cost optimization due to unused resources and one click jobs to remove them.

With a variety of resources being launched on your AWS cloud infrastructure on a daily basis, it is inevitable that there would be some resources that are left unused or under-used.

Use AWS Reserved Instances

One of the best ways to save cost is to reserve the instances (EC2, RDS, Redshift etc) to reduce costs from 20% to 60%. We have outlined the AWS RI plans with recent changes in our earlier blog post. A good RI planning can help you save significant AWS cost. Botmetric has a Reserved Instance planner to help you plan your AWS RI purchases. We strongly recommend businesses to evaluate their RI utilization and purchases on a monthly basis if you have growing usage of AWS cloud inside your company.

Having a check on unused RI reservations can also save your cost. You must check if your unused reservations can be modified correspond to some of your running instances. Availability zone and the network type (VPC/non-VPC) can be modified for an existing reservation.

Caching and Compression for CloudFront Content

From Botmetric’s cost analytics for hundreds of AWS accounts, we have noticed a pattern of modest to huge CloudFront spends. Most of the times, customers would have forgotten to enable caching for CloudFront resources either at origin servers like Web or S3. You should enable compression for content like CSS, Javascript, Text Files, JSON, HTML etc to ensure cost savings on bandwidth. Also, frequently deploy production changes and often forget to enable caching & compression for static resources, dynamic content like text/HTML/JSON etc. We recommend you to have post-deploy job as a part of your release automation to ensure client side caching, server-side compression etc are enabled for your application and resources.

To sum up, you need to have a good review process and timely alert internally along with smart tools ranging from AWS Trusted Advisor to Botmetric etc for effective cost managing and savings.

Botmetric’s comprehensive cost audit and analysis provides the list of unused and under-used resources. It also lists various cost optimization opportunities on your AWS cloud account. It also has a smart Reserved Instance planner to help you plan your reservations and alert you on unused reservations.

Get hold of your AWS cost spend now! TRY BOTMETRIC FOR FREE.