Performance Optimization

One of the pain points that DevOps teams suffer is too many knobs to tune for performance improvement. Trial and error is not a reliable way to find the right threshold of certain metric for autoscaling because it is a time-consuming process and is mostly likely resulted in over-scaling with wasted resources or under-scaling that does not meet performance goals when a peak load occurs.

Using intelligent autoscaling with application workload predictions, learns the capacity of each pod and scales the number of pods at the right time to meet the workload demands. The Applicaion-aware HPA provides a simple and cost-effective way for autoscaling application containers and achieving performance goals with the right number of resources.

Machine Learning-based HPA’s machine-learning based intelligent autoscaling eliminates the need for manual tuning/experimenting of different metric thresholds for best autoscaling result.
Just-in-Time Autoscaling
Continuous and real-time inputs of workload and performance metrics help to learn dynamic workload patterns and container capacity. This allows to autoscale the right number of replicas at the right time to meet workload demands.
Application-aware Autoscaling’s automatic scaling of Kubernetes application containers is based on application-specific workload and KPI metrics. This achieves autoscaling according to real application workload with performance target in mind.

Red Hat has put countless hours into curating our partner ecosystem and this includes Cloud Cost Management companies to help our customers manage their day 2 operations and onward.

Companies like ProphetStor run on our Red Hat OpenShift Container Platform to provide performance, cost and efficiency management at the workload level.

Cody Richard
Global Partner Solutions Architect, Red Hat

Red Hat logo