Armakuni’s Cloud Optimisation Journey with Karpenter
We’re thrilled to share that Armakuni has been recognized on the AWS Partner Network (APN). This opportunity reflects our dedication to providing top-tier cloud solutions and highlights our position as an AWS Premier Tier Consulting Partner. The APN blog is known for showcasing the most innovative work from AWS partners, and our inclusion here speaks to the quality and impact of the solutions we deliver.
Improving workload management with Karpenter
We recently helped a client tackle significant operational challenges by utilising Karpenter, an open-source autoscaler. This tool automatically adjusts computing resources based on current demand, helping us address our client’s pain points efficiently.
Client’s challenges
Our client, who operates a codeless AI platform for automating marketing decisions through machine learning, was struggling with inefficient cluster management. Their existing auto scaling solution was slow, sometimes taking up to 120 seconds to respond to changes in workload. This delay led to bottlenecks and higher operational costs.
Why we chose Karpenter
Karpenter’s ability to scale resources dynamically and in real-time made it the ideal solution. By integrating it into Amazon Elastic Kubernetes Service (EKS), we enabled more responsive autoscaling that closely monitors the Kubernetes cluster. This approach allowed us to fine-tune resource usage without increasing costs.
The results
After implementing Karpenter, we achieved:
- A 57% reduction in monthly EC2 costs, with expenses dropping from $4,200-$4,400 to $2,400-$2,600
- Improved infrastructure resilience by utilising multiple spot instance families (cheaper, excess cloud resources)
- Zero downtime during peak traffic due to better workload management
Key considerations when using Karpenter
If you’re considering Karpenter for managing cloud workloads, it’s important to keep a few things in mind:
- Stable Kubernetes cluster: Your Kubernetes environment needs to be robust, as it plays a critical role in automating the deployment and management of containerized applications.
- Reliable monitoring tools: Monitoring is crucial for tracking performance, identifying issues early, and optimising resource use.
- Thorough testing environment: Set up a detailed testing environment before deploying Karpenter to fine-tune the autoscaler and address any potential bottlenecks.
What this means for Armakuni and your business
Our solution not only boosted performance and reduced costs for our client but also strengthened their cloud infrastructure. This recognition on the APN underscores Armakuni’s expertise in cloud-native development, DevOps, and AI-driven solutions. We’re committed to helping businesses like yours navigate and optimise their cloud journeys.
You can learn more about our approach and how we can support your organisation by checking out the full details on the APN.
Meet the speakers
View all insights
Discover how Zero-ETL simplifies real-time data integration between Aurora and Redshift, eliminating manual pipelines and enhancing agility for faster insights and cost-effective data management.