Spotinst Ocean ensures 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.
However, estimating the proper amount of CPU and Memory when assigning resource requests to pods is also a challenge which teams are facing when designing a Kubernetes cluster.
To address this challenge and create an even more resource-efficient Kubernetes clusters, Spotinst had implemented a Vertical Pod Autoscaling (VPA) suggestions mechanism, as an integrated part of the Ocean solution.
Spotinst Ocean, empowered by VPA suggestions, monitors the Kubernetes Pods in real-time and aims to assist with recommendations for adjusting the resource requirements each deployment.
With VPA Suggestions, Ocean continuously monitors the cluster’s CPU and Memory usage and provides notifications when a deployment consumes significantly more\less resources than it is assigned.
Applying those notifications will help utilizing resources in the cluster in a more precise manner, and lower chances cluster issues as an outcome of under\over resource utilization on the other hand.