Spot EC2, EKS Node Groups, ECS, ECS Fargate, and ASGs to save avg. 60% on AWS bill

Systematically growing, cloud spending annoys budget-minded organizational pros, despite cloud experts trying to handle the cloud waste challenges for the sixth year, promoting various saving opportunities. One of them is converting instances into spots, which means purchasing already bought but not used capacity for a whooping discount of up to 90%. It is impressive. So, let’s discuss these magical and, at the same time, tricky spots. What are they? How to convert mission-critical workloads into spots confidently? And what is multi-service spotting?

Spots: Saving opportunity we all should know of and resort to

Amazon was the first to come up with spots. Users who buy AWS resources use them less than 100% since planning for the exact use from the start is challenging. As a result, a certain amount of idle resources happened, and Amazon figured out how to resell them at an impressive discount of up to 90%.

Spot Instances can arrange the most considerable discount. However, Flexera’s report reveals that less than 30% of AWS users are converting their instances to spots, uncovering that most companies do not risk taking full advantage of this opportunity. Why? Spotted resources are disconnected and taken back in the shortest possible time, from 2 minutes to as little as 30 seconds, leaving little time to do something about it. And, of course, if organizations include spotting to their infrastructure operation plan, then do it very carefully, and concerning stateless, error-liberal workloads or processes that can be painlessly stopped and restarted at any time.

On the other hand, we can naturally assume that the more resources we convert into spots, the more we save. Converting your mission-critical workload resources into spots is the most efficient method to reduce cloud costs. So, here is the question: how can you do it confidently, without fear of abrupt terminations? Without spot management strategy and well-architected automation, in no way.

Spot EC2, EKS Node Groups, ECS, Fargate with ECS, and ASGs with Uniskai’s spot management strategy

AWS experts propose maintaining workload availability by proactively increasing your Autoscaling group (ASG). If you’re planning to spot an EC2-only, it must be attached to ASG. What if an Autoscaling Group is not an option? What about other resources except for pure AWS EC2?

Profisea Labs experts created Uniskai to enable users to visualize and optimize their cloud infrastructure and take control of their cloud. In terms of savings, Profisea Labs’ pros combine all cloud cost optimization best practices into unique saving strategy where Uniskai reduces costs with a patented spot management technology that includes:

  • Automated spot pricing observation to make a bid considering the default maximum price
  • Fast spot converting to switch EC2, EC2s inside EKS Node Groups, ECS, Autoscaling Groups and Fargate with ECS into spots

Short newsletter: Here is the tab of all services we will discuss today (Amazon ECS, EKS, Fargate ECS, AGS) with brief information about them. If you need to read and refresh some info, do it. If all is clear, skip it.

  • Spot Instance uninterrupted management allows Uniskai’s systems to promptly handle the resource termination, applying redeploying reactions
  • Spot Instances’ intelligent scheduling to set up automated hibernation for spotted resources to save cloud costs the most.

Multi-service Spotting rocks

Let's discuss spotting EC2, ASGs, EKS Node Groups, ECS, ECS on EC2 and Fargate.

Cloud-native spot strategies and spot scenarios proposed by third-party solutions mainly provide spotting for container/cluster-separated computing-powered machines, like Amazon EC2. However, practical as it is, you can save an impressive sum of cloud costs when easily converting the highest number of group, container and cluster-packed instances as a part of ECS, EKS, and ASGs, with ECS Fargate full-fledged converting.

Simple example: Here is one of the AWS accounts attached to Uniskai, where our platform’s algorithms detected only spot-related opportunities to save about $400 a month.

So, if we look carefully, we will see that here we have resources listed, and this list goes far beyond pure EC2 instances. Therefore, if EC2s-only were converted into spots, the saving would be less than $200.

How simple is this Math? It is much more pleasant to save $400 than $200. Right? My point is that with Uniskai, you have more extended saving opportunities to convert many expensive cloud resources into spots without fearing interrupting your mission-critical workloads.

Final thoughts

Cloud providers give free resources at a discounted price. One condition: sudden instance closure can pop up at any moment. And the best and only way to handle spot instance sudden interruptions is via a well-architected spot management strategy. So who can create and implement this strategy? It’s a good question. Thus, to help with the answer, let’s put the pieces together:

  • Converting resources with unused capacity (even for business-critical workloads) into spots efficiently reduces cloud costs.
  • Confidence in spot converting can be possible and compelling when supported by a well-architected spot management strategy. It enables a one-minute-long reincarnation of the instance with all previously downloaded data and credentials saved.
  • The most impressive cloud savings can be achieved due to multi-service spotting involving various resources, from container-Kubernetes core systems to Autoscaling groups.

Uniskai by Profisea Labs effectively manages spots and maintains the required availability owning to its proactive steps before any termination takes effect. The innovative approach of being prepared for spotted multi-services abrupt ending combined with spot pricing monitoring resulted in a powerful Spot Instance management strategy by Uniskai that you can utilize to avg. 60% of cloud costs.