How to Monitor and Control AWS Costs

aws cost
As the leading cloud service provider, AWS has attracted a significant number of organizations prioritizing on cost efficiency through the cloud. Although this has largely been effective for most companies, there are still a couple that are operating on a higher budget than they earlier anticipated. They may have successfully reduced their overall costs- but they’re still yet to develop strategies to further minimize them. Such organizations are yet to fully realize the costs benefits of adopting cloud solutions.

If you’re part of such a company, you may be pleased to learn that all hope is not lost. AWS has a couple of strategies which could be employed to reduce your computing costs and subsequently control them effectually- but still maintain optimal computing performance. Here are some of the most efficacious tips to monitor and control AWS costs:

Keeping Up With New Instance Types

Amazon are famous for the attention they give their users. They consistently review their user’s comments and consequently improve their products to release new instance types according to their respective customer needs.  Of course it’s inadvisable to blindly adopt new instance types- but definitely worth it if potential cost savings outweigh the cost of adopting new solutions. It’s therefore prudent to review each new instance type according to its potential cost savings and overall computing benefits before considering it in your cloud strategy.

Although some don’t impact costs, a significant number of new instance types have a potential to further reduce your cloud costs. In 2014 for example, Amazon released new T2 instances which create “CPU Credits” during low-processing periods to automatically utilize during busy periods. This low-cost stable processing approach has proven to be particularly ideal for applications with rare spikes like development tools and small databases.

Leveraging Spot Instances

To cheaply operate and run background processes, you should consider leveraging spot instances. They allow you to dictate the maximum amount you’d be comfortable spending on the instances, consequently facilitating effective price control particularly during off-peak hours.

The main disadvantage of this feature however, is the fact that your instances may be prematurely terminated when the price is exceeded. Your processes could be stopped even when the jobs are 90% complete, resulting in cost and resource wastage since you’ll be compelled to restart the processes from scratch. The best strategy to prevent this is building an architecture with dynamically changing bid prices which don’t exceed on-demand prices. That means setting up a spot instance only auto-scaling group- CloudWatch will regulate the entire process by scaling up the groups within request parameters when price meets a bid. This setup should further be combined with a second on-demand auto-scaling group, with an ELB serving as a link between the two groups. Through it, requests will be efficiently shifted between on-demand and spot groups.

Reserved over On-Demand Instances

In a bid to reduce their overall cloud computing costs, many businesses have preferred reserved instances over on-demand instances. They are acquired for a period of one to three years, through an upfront or partial payment system. In some cases, users have managed to reduce their cloud computing costs by more than 50% by shifting from on-demand to reserved instances.

Although it may sound promising, it’s not all a bed of roses- there are a couple of complications which come with this option. For instance, it’s fairly difficult to correctly predict the amount of instances you’ll need over the next couple of years, especially if you’ve not been an avid user of AWS. Only the most experienced cloud users can accurately do so.

Adopting Monitoring Tools

The most fundamental step in cloud cost assessment and subsequent control is collecting accurate information on your regular costs vis-à-vis resource usage. Informed strategies can only be drawn up from comprehensive facts on regular cloud usage. Amazon has simplified this process by granting access to a wide range of tools that you could use to collect historical and current information on your cloud usage. Some of these third party applications additionally provide cost saving tips according to user behavior and cloud architecture. Two of the most prominent ones include;

  • Trusted Advisor: Although it provides fault tolerant, security and availability recommendations, Trusted Advisor’s main function is to recommend tools and opportunities which would significantly reduce your cloud spending. It monitors user CPU usage and has the capability of reducing your cloud costs by 20-50%
  • Cloudcheckr: This is the ideal tool if you need to track any unusual costs which may appear on your overall AWS bill. It uses sophisticated tools to assess your cloud environment and report results in detailed billing analysis and cost maps to give you a vivid picture of the underlying cloud operations and related costs. Consequently, standard policies are enforced and developers are alerted in cases resources are run outside that configuration.

In the year 2014, Amazon added Cost Explorer in the list to help users further monitor and manage their cloud computing costs. It keeps track of and displays regular reports on daily and monthly spend, monthly cost by linked account, and monthly cost by service.

By collectively combining these features in assessing, monitoring and controlling your AWS cloud costs, you’ll be in a better position to make significant savings which may even run into millions of shillings, depending on your company size. Such amounts may alternatively be invested in other business departments to improve your overall operations, service delivery and ultimately trigger exponential business growth.

Author: Davis Porter

Image Courtesy: Natara, and Amazon