The dark side of Kubernetes? Light it up with application-aware optimization

There is a “dark side” to Kubernetes that makes it difficult to ensure the desired performance and resilience of cloud-native applications, while also keeping their costs under control.

Indeed, the combined effect of Kubernetes resource management mechanisms and application runtime heuristics may cause serious performance and resilience risks.

• How to best align the sizing of the container with the resource requirements of the applications running on it?
• How to avoid CPU throttling impacting response time?
• What is the right heap size and Garbage Collection type to minimize cost and avoid out-of-memory (OOM) issues?

On the other hand, there are also significant potential improvements, both in terms of performance and efficiency, that can be achieved by properly tuning Kubernetes and application runtime (e.g. JVM, Golang) configuration settings.

In this webinar, we illustrate how the Akamas AI-powered optimizations platform addresses these challenges by making it possible to set the optimization goals (e.g. cost reduction) and constraints (e.g. performance SLOs) and get recommendations on how to adjust configuration settings dynamically under the varying workload.