Kubernetes automation platform startup PerfectScale launches first channel program

PerfectScale’s platform, which uses machine learning to remove the bloat that developers around Kubernetes, works with all Kubernetes systems, and the program is designed to assist a broad range of partner types.

Today, PerfectScale, a Kubernetes optimization startup based in Morrisville in North Carolina’s Research Triangle, is launching their first channel program. Aimed at a broad variety of partner types, including cloud consultants and MSPs as well as resellers and referral partners, the PerfectScale program contains some differentiating features, as well as providing a platform which will improve the effectiveness of all the major Kubernetes vendors in the market.

PerfectScale, which emerged from stealth this spring following its beta, was founded by three individuals, one of whom spent many years in DevOps environments managing Kubernetes environments.

“The challenge in these environments was to ensure that Kubernetes was running on optimal machines and that it was not overprovisioned,” said Amichai Lesser, VP of Business Development at PerfectScale. “Kubernetes is a very dynamic environment. If it is not set up correctly, it can quickly become bloated. The typical Kubernetes cluster has three times the resources that it needs.”

Typically, this situation has been addressed manually, in a process that is long and not especially productive.

Amichai Lesser, VP of Business Development at PerfectScale

“If you fix it by hand, and define resources manually, it is tediious and time consuming,” Lesser said. “We listen to what’s going on in each cluster for each job – bearing in mind that there could be hundreds or thousands of jobs in each cluster. We automatically pinpoint the mismatched jobs which have too many resources, as well as the occasional job which has inadequate resources, which can cause it to fail.”

Lesser said that this has not been done before because it was not a trivial problem to solve.

“It involves very sophisticated machine learning algorithms,” he indicated. “The  mismatches themselves are relatively easy to pick up. It’s understanding the impact that is more important. You could have a job that is  1000% overprovisioned, but which only runs for five minutes a month, which would be much less problematic than a job that is overprovisioned 20%, but which runs 24/7.”

While the PerfectScale platform has obvious benefits for enterprises with more clusters, to date, most of their sales have been to born in-the-cloud technology companies.

“We have close to 100 customers, from startups to scaleups to enterprises,” Lesser said. “Typically, they are born in-the-cloud companies that started with Kubernetes from day one. These technology companies don’t have to be big to get good value from us, especially as Kubernetes bloat starts early. We help them save 30-40% of the amount that they spend. We have also started to sell to enterprises as well.”

PerfectScale doesn’t compete against the hyperscalers who sell Kubernetes, or with other types of Kubernetes providers. Instead, it optimizes all of them.

“We have purely a cooperative relationship with them,” Lesser noted. “We help all the Kubernetes relationships, and are agnostic from a technology perspective. We support them all.”

To date, PerfectScale has used partners in geos like EMEA and APAC where they have no direct sales people, and in cases where the partner brought in the customer.

“Everything we did so far was small scale,” Lesser said. “Many companies our size don’t invest in a partner program this early, but there is a very active ecosystem of cloud consulting partners around Kubernetes. So we have a consulting partner program as well as one for MSPs. These are usually things that come in when a company is more mature, and after they start with refer and resell – but these are a dynamic part of the ecosystem.” They have a self-service option as well. All of these have a single tier.

They don’t work with larger systems integrators – not yet anyway.

“We are too small for the big SIs to build a practice around us at this stage, as the  – market is still in its infancy,” Lesser said.

Partners in the program can go to market through a wide variety of collaboration formats including co-sell, re-sell, managed service and the bundling of PerfectScale inside consulting service offerings.

“These are not exclusive to specific types of partners,” Lesser stated. “Resellers for example, could choose to resell or refer, depending on the customer. We see the model as a flexible hybrid one.”

Lesser indicated that in the consulting track, they expect to see dozens of partners.

“On the reseller side, we will be a bit more selective, and will limit the number of resellers at first in key geos,” he said.

Tools include the ability to register leads on a portal that is integrated with a CLM system, although Lesser said that am upgraded version will be coming.

In-depth sales training and engineer-led technical training are provided.

“We have consulting certifications now,” Lesser said. “We don’t have sales certifications yet, but they are not that far out, likely mid-2024. It’s not hard to master this in any event.”

Exclusive branding kits, sales, marketing and educational collaterals, and the ability to sell directly through the AWS Marketplace are all available.

Lesser also highlighted a couple of differentiating features in the program.

“We provide health checks to partners for free,” he said. “Any partner can offer a Kubernetes health check and in one week get a health check report on bloat and optimization opportunities. Whether they want to charge the customer for this is up to them.”

The second feature is a dedicated PerfectScale SaaS instance for MSPs and CSPs.

“This provides the ability to get your own instance and onboard customers with a Parter SaaS.” Lesser said.