Sign up to our Newsletter
Containers are the foundation of today’s hyper-agile DevOps landscape. Developers can enjoy the reliability of a single environment to run code, with no unexpected hiccups or changes between testing and production. Segmented applications benefit from faster and more efficient delivery, as each functionality evolves via its own isolated lifecycle.
As powerful as the foundation that containerization offers, once deployed to Kubernetes, the most difficult challenge is ensuring every container has the resources it needs at any given time.
How One AI-Driven Media Platform Cut EBS Costs for AWS ASGs by 48%
In this blog post, we are going to cover how we’ve managed to reduce costs on Kubernetes by 91% for a major FinTech firm within just a few hours, while optimizing performance.
Even Google Recognizes Kubernetes Can Be Complicated
Kubernetes has become the de facto standard for container orchestration, making it a popular choice for businesses of all sizes that want to streamline their operations and scale their applications with ease.
Despite its benefits, the process of managing resources on Kubernetes can be incredibly costly and time consuming. Even with enough knowledge, the many moving parts – alongside a lack of graphical user interface – can present a major challenge to set up and configuration. As the solution’s inventor and most prominent advocate, Google has previously issued a statement on the widespread difficulties faced by enterprises. Within this, they acknowledge that they’ve seen many companies enthusiastically embrace Kubernetes before running “headlong into difficulty”. The complexity of its resource management causes operational overhead and critical problems – from CPU throttling to over-provisioning.
As FinOps continues to evolve and push for increasingly real-time decision making, it’s vital that the tools on hand support lean and optimized cloud spending. The sheer complexity of Kubernetes architecture sees DevOps teams consistently waste significant time and energy on never-ending manual configurations.
One major FinTech company was faced with the daunting price tag of $816,000 per year on K8s resource management. After spending a lot of time and resources on managing, analyzing, and right-sizing their workload, they have turned to us for assistance.
Innovative FinOps Technology Ensured That Resource Requests and Limits Were Fully Optimized
Led by a team of seasoned engineers and architects, we knew that this firm would benefit immensely from Automated Kubernetes Optimization Platform.
With rapid installation and continued support, this cutting-edge approach to K8s visibility lent the FinTechOps team real-time insights into the organization’s wider K8s cost structure, alongside resource consumption anomalies and alerts. With an on-the-ground view of what resources each system actually needs, the solution’s autonomous pod rightsizing was able to kick in. These in-place updates ensured that every request – and the limits the firm paid for – were fully optimized and streamlined to our client’s actual usage.
Breakdown: GlobalDots’ FinOps Solutions
The difference was instant and phenomenal: whereas resource management was previously a significant revenue drain, the annual cost post-optimization now sits at $72,000 instead of $816,000!
The firm was having between 70 to 120 worker nodes running to serve their customers during peak and off-peak times. With the deployment of the automated K8s platform on their K8s clusters, the number of nodes actively requiring management has fallen just as sharply.
We’ve seen exactly which workloads were not configured wisely and consuming the wrong amount of resources and replicas, leading to the inefficiency of our customer’s clusters. With the autonomous K8s optimization solution, it was super quick and simple to configure the right amount of resources for each workload, using the advanced algorithm and performance capacity management.
The average rate of savings across all clusters is now projected to be at a staggering 91%. This success story is only one of thousands that we’re deeply proud of. GlobalDots’ multi-vendor approach has allowed leaders such as SentinelOne, Gong, and Playtika, to optimize their cloud costs by introducing cutting-edge solutions. To discuss how your cloud architecture can benefit from GlobalDots‘ innovation hunting skills, get in touch today.
