groundcover Acquires Wand Platform to Optimize Kubernetes Clusters

TL;DR — Key Takeaways

– groundcover Acquires Wand: The deal brings automated Kubernetes resource optimization to groundcover’s observability platform.

– Smarter Resource Allocation: Wand dynamically adjusts CPU and memory allocations to reduce infrastructure waste.

– Lower Observability Costs: groundcover plans to use Wand to improve the efficiency of its BYOC data plane.

groundcover today revealed it has acquired a platform for optimizing consumption of infrastructure resources running on Kubernetes clusters that was developed by Wand. Terms of the deal were not disclosed.

Company CEO Shahar Azulay said the acquisition of this asset represents a significant step toward closing the gap that currently exists between observability and IT infrastructure automation.

The Wand platform continuously analyzes workload behavior across an entire cluster versus focusing only on a single workload at a time. It then automatically adjusts CPU and memory allocation on live infrastructure as demand changes to reduce overprovisioning of IT infrastructure. Additionally, Wand can be installed using a familiar Helm chart, with IT teams able to track changes made to the cluster.

The groundcover platform, meanwhile, captures telemetry data from production environments using the OpenTelemetry frameworks and extended Berkeley Packet Filter (eBPF) in a way that doesn’t require IT teams to first instrument every application.

The Wand platform will then make it possible for IT teams managing Kubernetes clusters to automatically apply the insights surfaced by the observability tools provided by groundcover, said Azulay. The overall goal is to further adoption of autonomous workflows by providing the context artificial intelligence (AI) agents will need to reliably and safely manage operations autonomously, he added.

Additionally, groundcover will use Wand to make its own BYOC data plane more efficient and resilient. Because groundcover runs inside customer cloud environments, the resources it consumes appear directly on customer cloud bills. Making storage and processing more efficient lowers that cost, which in turn lets customers economically retain more telemetry at full fidelity, giving engineers and agents more context to work with.

In general, there is a lot more attention being paid to optimizing Kubernetes resource consumption for two reasons. The first is that rising processor and memory prices are increasing IT infrastructure costs, an issue that is further exacerbated by a shortage of servers because much of the manufacturing capacity for processors and memory is being allocated to providers of various AI services.

Secondly, IT teams are starting to deploy more AI applications on Kubernetes clusters. The graphics processing units (GPUs) that are relied on today to drive those applications are also in short supply, leading to higher costs. Unfortunately, utilization rates of GPUs in servers are often in the single digits. IT teams are now squarely focused on increasing those utilization rates as part of an effort to optimally deploy as many AI applications as possible, noted Azulay.

The challenge is that most IT teams today are trying to manually tune a Kubernetes cluster that requires them to set CPU and memory requests prior to an application being deployed. The end result of that guesswork is often a lot of wasted IT infrastructure resources that could be better optimized using an IT automation framework to manage both horizontal and vertical scaling across the entire cluster, said Azulay.

In the absence of that capability, the autoscaler installed on a Kubernetes cluster may add additional replicas to ensure a service level is maintained without regard for how those resource allocations might impact other applications running on that cluster.

Hopefully, there will come a day when any mere mortal IT administrator can successfully manage a fleet of Kubernetes clusters. In the meantime, however, it’s apparent that the first step toward achieving that goal will be to rely more on IT automation frameworks augmented by AI.