Fresh off raising $4.15 million in seed funding, Replex has made available a namesake governance and cost management platform designed for Kubernetes environments.
Company CEO Patrick Kirchhoff says that although IT service management tools that provide visibility into legacy IT environments are plentiful, organizations that have adopted Kubernetes don’t have anywhere near the same level of insight into how and when Kubernetes clusters are being deployed. That usually results in a lot of extra costs whenever a development team decides to spin up its own Kubernetes cluster.
Replex, he says, reduces those costs providing the insights IT organizations require to right size clusters, pods and container sizes for optimal utilization. Based on real-time and historical utilization trends, Replex also forecasts usage, manages chargebacks and plans future infrastructure capacity requirements, which can result in as much as a 30 percent savings on IT costs, Kirchhoff claims.
Kirchhoff says because it’s possible for developers to programmatically spin up a Kubernetes cluster almost anywhere, IT leaders and financial executives want to understand the cost implications of deploying applications running on top of Kubernetes in one platform versus another. It may be easier to spin up a Kubernetes cluster on a public cloud, but over time many organizations will discover that long-running cloud-native applications are less costly to deploy in an on-premises IT environment. Conversely, organizations that have committed to migrating all their applications to the cloud can employ the optimization engine that Replex developed to better understand their true costs, he says.
Adoption of Kubernetes within most enterprise IT environments is still relatively nascent, so it’s still relatively early days as far as accountability from a financial perspective. Most of the early adopters of Kubernetes are far more interested in understanding how well Kubernetes enables their organizations to become more agile. However, a day of financial reckoning is coming. Many organizations are already repatriating some of the monolithic applications they built and deployed on public clouds into on-premises IT environments because use of those applications is turning out to be higher than anticipated. Cloud-native applications built on top of Kubernetes are designed to dynamically scale out as needed, which also can lead to sudden unexpected increases in billing from the cloud service provider hosting those applications.
In fact, Kirchhoff notes that lack of visibility into costs may limit Kubernetes adoption in the short term. Most enterprise IT organizations have spun up at least one Kubernetes cluster. Most of them, however, are not going to deploy Kubernetes at scale in production environments until they really understand the total cost of Kubernetes ownership. Keeping track of those costs using traditional IT tools has become next to impossible in this dynamic age of DevOps, which prizes rapid application deployment above all else.
It’s clear that in many ways Kubernetes can be too much of a good thing. The challenge now is finding a way to apply governance models without compromising the rate at which Kubernetes allows applications to be built and deployed at scale.