Pepperdata announced that its product portfolio now provides autonomous optimization and observability for Spark applications running on Kubernetes.
Kubernetes is a key part of the modern hybrid, multi-cloud architecture in today's enterprises, notes Pepperdata. As big-data applications move from Spark on legacy systems to Spark on Kubernetes, the performance of these applications can change dramatically.
Pepperdata promises full-stack observability for Spark on Kubernetes, allowing developers to manually tune their applications while autonomously optimizing resources at run time. The combination of manual and autonomous tuning is necessary to deliver the best price and performance for these applications. Pepperdata uses machine learning across clusters, containers, pods, nodes, users and workflows to give you a complete understanding of your environment.
Unlike traditional infrastructure monitoring or manual tuning, which are limited in both scaling and speed, Pepperdata will automatically optimize Kubernetes resources while providing a correlated and granular understanding of the applications and infrastructure, they note, adding that observability provides actionable information to debug and understand complex applications, and autonomous optimization ensures that the compute resources are used efficiently.
Features include:
- Autonomous optimization of resources and workloads on Amazon EKS, HPE Ezmeral and Red Hat OpenShift
- Application and infrastructure observability for Spark on EKS, Ezmeral and OpenShift as well as YARN
- A self-service dashboard so developers can manually tune using recommendations for speed or resource utilization
- Detailed usage attribution for chargeback
"Kubernetes is becoming increasingly important for a unified IT infrastructure, both in the cloud and the data center. Spark is the number one big data application moving to the cloud, but Spark applications tend to be quite inefficient. Optimization is key to successful implementations," said Ash Munshi, CEO, Pepperdata. "We saw this early on with our customers, which is why we invested in the development of Spark on Kubernetes, together with Red Hat, Palantir and Google."
