Companies, especially those reliant on cloud operations, often struggle with resource optimization and configuration.

One of these examples is the automatic scaling of Kubernetes, a pain point that StormForge, which focuses on automated Kubernetes resource management, is working hard to alleviate for its customers.

“We started out running our machine learning workloads and moving them into Kubernetes,” said Patrick Bergstrom (pictured, left), chief technology officer of StormForge. “And then we weren’t quite sure how to correctly adjust and size our containers. And so our ML team got together and wrote an algorithm, and then we said, ‘Well, holy cow, that’s actually really useful. I wonder if other people would like that?’ And that’s where we got started.”

Bergstrom and Yasmin Rajabi (pictured, right), vice president of product management at StormForge spoke with theCUBE industry analysts Lisa Martin and Savannah Peterson at KubeCon + CloudNativeCon NA 2022, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed the tech behind Stormforge’s approach to K8s auto-scaling. (* Disclosure below.)

Building a business around a problem solved

StormForge started out as a machine learning shop before pivoting into automatic Kubernetes resource management, according to Rajabi. Having long embraced its new calling, the company considers input from its diverse user spectrum to develop and deploy new product features.

“For us, because we use machine learning, it’s a lot of building confidence with our users,” Rajabi explained. “So, making sure that they understand how we look at the data, how we come up with the recommendations and actually deploy those changes in their environment. There’s a lot of trust that needs to be built there.”

Kubernetes has both vertical and horizontal pod scalers. Where these two would crash upon contact, StormForge has found a way to harness them simultaneously to improve efficiency without sacrificing performance, according to Bergstrom.

“The big release we just announced is with our machine learning — we can now do both,” he said. “And so we can vertically scale your pods to the correct size and still allow you to enable the HPA. And we’ll make recommendations for your scaling points and your thresholds on the HPA.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of the KubeCon + CloudNativeCon NA 2022 event:

(* Disclosure: StormForge sponsored this segment of theCUBE. Neither StormForge nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE