SUSE acquires observability platform StackState

The open-source company SUSE announced the acquisition of the observability service provider StackState at its in-house exhibition SUSECON. StackState specializes in collecting metrics and logs in Kubernetes clusters, and thus supports administrators in troubleshooting.

Goal: Rancher Prime as an all-in-one solution

To provide users with a solution under one roof, SUSE wants to integrate the StackState platform into its Rancher Prime container management platform. During the keynote on the first morning of SUSECON, SUSE CEO Dirk-Pete van Leuween also announced that the source code of StackState would be published. In a blog post on the acquisition, he writes that Rancher Prime has been well-prepared for the integration of StackState so that users can quickly find their way around. CEO of StackState, Andreas Prins, told c't that the news of the new open source course had triggered a lot of enthusiasm among the development teams.

Although StackSate was previously a proprietary product, open standards such as eBPF and OpenTelemetry are firmly part of the platform's observability toolbox, Prins assured c't when asked. Anyone who wants to can also export data to operate their own dashboard based on Prometheus and Grafana, for example. There is no date yet for when Rancher Prime 3.1 with StackState integration will be generally available, but it is expected that the functions will be activated for existing Rancher Prime customers this year. Those using the community version of Rancher will be able to work with StackState via a SaaS offering.

AI focus for container management platforms

According to SUSE, StackState is intended to help customers keep their AI workloads under control, among other things. However, the developers want to add further functions to Rancher Prime in the upcoming version 3.1 to improve compatibility with AI workloads. These include virtual cluster provisioning, resource optimization for GPUs and the integration of KubeFlow into the application catalog ("Application Collection"), a machine learning toolkit for Kubernetes under the auspices of the CNCF (Cloud Native Computing Foundation).

Version 1.3.1 of the Kubernetes-based hypervisor Harvester is intended to improve compatibility with vGPU and ARM architectures, both of which are important resources for inference workloads. NeuVector Prime 5.4 is intended to display the results of pod and node scans directly in the Rancher Prime user interface and provide better protection against denial-of-service attacks. The developers have also integrated a new compliance reporting framework. (ndi)