Cloud Financial Management (FinOps)
Autoscaling applications
Amazon Web Services
Big data
Container infrastructure
Microsoft Azure
Reserved commitment management
Cloud services for MSPs
Google Cloud
Spot Overview
About NetApp
Elastigroup
Ocean
Ocean CD
Spot Connect
Eco
Cost Intelligence
Billing Engine
CloudCheckr
Ocean for Apache Spark
Spot Security
Case Studies
Resource Center
Documentation
News
Service Status
What Is CloudOps?
How to Operationalize FinOps
About Us
Contact Us
Careers
Managing production Kubernetes environments is indeed a challenging task, especially if you are motivated to empower cost reduction and cluster efficiency.
6 months ago, at AWS Re-invent 2018, we have announced Spotinst Ocean, our Serverless Kubernetes engine, which abstracts Kubernetes clusters from the underlying VM infrastructure and reduces cloud-compute costs by reliably leveraging Spot Instances.
Spotinst Ocean ensures that Pods are placed on the best possible VM that fits the Pods’ resource requirements, along with other scheduling constraints such as taints, tolerations, and affinity rules.
Throughout the Ocean journey, we quickly realized that engineering teams are in a constant struggle of estimating the real and correct amount of CPU and Memory when assigning resource requests to their Kubernetes pods.
Here at Spotinst, our obsession for cost-oriented Kubernetes clusters is what drove us into adding Vertical Pod Autoscaling (VPA) suggestions as an integrated part of the solution.
Starting today, Spotinst Ocean, empowered by VPA suggestions, monitors your Kubernetes Pods in real-time and assists you in adjusting their resource requirements, in order to achieve a more efficient Ocean cluster, which will significantly lower your operational costs.
With VPA Suggestions, Spotinst Ocean continuously monitors the cluster’s CPU and Memory usage and will notify you when a Pod repeatedly continues to utilize fewer resources than those assigned to it.
On the other hand, in cases which a Pod is over utilizing its assigned resources, you can increase the CPU and Memory of the Pod in order to avoid Pod termination as an outcome of OOM (out of memory) and lack of CPU.
The notification may be used to modify the deployment size so it will request only the amount of CPU and Memory that it actually consumes, and not the uncertain amount the developer has initially assigned to it, which was usually defined based on a rough assumption.
This new feature is enabled for all AWS Ocean clusters – Jump to your Spotinst Ocean console and get ready to sail to a more efficient and hassle-free future!
Keep reading about Ocean VPA Suggestions in our blog post