Summary

• Sedai helped a major Kubernetes user save $500,000 in annual cloud costs within 60 days through rightsizing in Dev/Test environments.
• The optimization focused on 1,400 Kubernetes services, adjusting CPU, memory, pod counts, and the configuration of instances and node groupings.
• Prior to adopting Sedai's autonomous optimization, manually optimizing each service was not cost-effective due to the minor individual costs.
• The use of Sedai's AI-driven technology marked a shift from manual processes to efficient, autonomous operations.

Problem

The customer faced significant challenges in managing their Kubernetes environments, stemming from the complexity of numerous services, rapid business growth, and the demanding requirements of their CI/CD pipeline. The engineering teams were strained, and there was a growing concern about the potential for human error, burnout, and the inefficiency of cloud resource usage. Manually optimizing the myriad of microservices, each with relatively small individual spend, was proving uneconomical and unsustainable.

Solution

In response to these challenges, the company decided to adopt Sedai's autonomous optimization solution. Sedai's technology was integrated within the customer's Google Kubernetes Engine (GKE) environments, utilizing a "bring your own cloud" deployment model to meet their specific security and access requirements. Sedai was granted permissions to map the account topology and analyze behavior patterns, which is crucial for effective Kubernetes cost optimization. Sedai's goals were focused strictly on cost optimization, aiming to find the most efficient use of Kubernetes resources while maintaining current performance levels. Sedai applied AI techniques to determine the optimal workload and cluster configuration.

Results

The implementation of Sedai's technology led to substantial cost savings in less than 90 days. By rightsizing the Kubernetes resources, primarily through adjusting the configurations at both the workload and cluster levels, the company saved $500,000 on an annual runrate basis. This was achieved by optimizing 1,400 Kubernetes services. With an average saving of around $400/service this would have been uneconomic for engineers to complete manually.

Looking Ahead

Looking forward, the company plans to expand their optimization efforts beyond the initial Dev/Test environments. They aim to include additional clusters and leverage more advanced features of Sedai for unique settings, like machine learning-based workloads. The move towards more automated and eventually autonomous operations continues, with a vision to extend these optimizations into production environments. The company is also considering broader Kubernetes infrastructure cost strategies, such as using spot instances and scheduled shutdowns, to further enhance their cost efficiency.

‍

‍