groundcover Opens Beta for Automated Kubernetes Resource Tuning

groundcover opened a beta Nov. 3 for software that adjusts CPU and memory requests for Kubernetes workloads using production usage data. The feature, Autonomous Infrastructure, draws on technology from Wand, a Kubernetes resource optimization company groundcover acquired in September.

Kubernetes uses resource requests when deciding whether a workload can fit on a node. A workload’s actual CPU or memory use can exceed its request when resources are available. Groundcover’s beta changes requests while leaving Kubernetes limits unchanged.

The feature is off by default. Administrators choose which deployments, namespaces or clusters it can manage, then set minimum and maximum request levels. Before raising a request, the software checks the capacity available on the node. It also records policy assignments in an audit log.

A dashboard compares requested CPU and memory with actual usage, flags workloads that may be under-provisioned and displays estimated monthly savings. The savings figures are estimates. Groundcover customers running its Kubernetes sensor can join the beta at no additional cost.

Supported sensor versions can display the information without a new sensor installation or a waiting period for telemetry to accumulate. The feature currently supports Kubernetes Deployments, including injected and native sidecars. Other workload types are on the roadmap.

Groundcover acquired Wand on Sept. 24. Wand’s founding team and employees were set to join groundcover, while the acquisition announcement did not disclose deal terms.

Groundcover is recruiting design partners and planned demonstrations at KubeCon + CloudNativeCon North America in Salt Lake City, Nov. 9-12.