Managing Kubernetes across AWS, Azure, on-premises data centers, and edge locations simultaneously? That’s the infrastructure nightmare keeping platform engineers up at night. Mirantis thinks they have an answer with k0rdent v1.2.0, their open-source “super control plane” that just added highly requested OpenStack support alongside improved ARM64 compatibility and a complete observability overhaul.
The fundamental challenge k0rdent addresses isn’t new, but it’s gotten significantly worse with AI workloads. Modern applications now spread inference processing across edge devices, training pipelines in private data centers, and data storage in public clouds. Each environment needs its own Kubernetes cluster, and suddenly you’re managing dozens of separate control planes with different configurations, versions, and tooling. k0rdent aims to centralize that chaos through templated, declarative management.
Three Components Handle Different Infrastructure Layers
K0rdent breaks down cluster management into three distinct modules that work together but handle separate concerns.
K0rdent Cluster Manager (KCM) takes care of the boring but critical stuff: cluster lifecycle management, upgrades, and scaling operations. Instead of writing custom Terraform scripts for each environment, KCM uses plain YAML templates that integrate with GitOps workflows.
K0rdent State Manager (KSM) deploys and manages the services running on those clusters. Think Istio for service mesh, Flux for continuous delivery, or cert-manager for TLS certificates. KSM uses the same declarative template approach, so teams can maintain consistency across all their environments without copy-pasting configuration files.
K0rdent Observability & FinOps (KOF) is where things get interesting for teams tracking cloud costs. This module generates metrics, logs, and dashboards through integrations with VictoriaMetrics and OpenCost, giving visibility into exactly what each cluster costs to run.
OpenStack Template Arrives After Community Workaround
The headline feature in v1.2.0 is OpenStack Hosted Control Plane support, something the community had been requesting for months. The delay wasn’t Mirantis dragging their feet—it was a bug in Cluster API Provider OpenStack (CAPO) that prevented proper implementation.
Rather than wait for an upstream fix, the k0rdent team built a workaround that bypasses CAPO entirely for network infrastructure creation. The tradeoff? Administrators now need to manually configure networks, subnets, routers, and load balancers before deploying clusters. Not ideal, but it gets OpenStack users unblocked while the CAPO issues get resolved.
Azure users get more flexibility in this release too. The updated templates now accept custom image sources beyond Microsoft’s default marketplace options, useful for organizations with security requirements around base images or those running highly customized distributions.
Observability Module Gets OpenTelemetry Upgrade
The most significant technical change in v1.2.0 happens under the hood in the observability stack. K0rdent has migrated from Victoria Metrics collector to OpenTelemetry, the industry-standard observability framework that’s become the de facto choice for cloud-native applications.
This isn’t just swapping one collector for another. The new implementation uses four specialized collectors that each handle different data sources: kube cluster collectors pull cluster-level statistics, node daemon collectors gather host metrics, a k0s components collector polls etcd and controller manager directly, and a syslog collector with Grok pattern extraction processes unstructured log data.
Better metrics labeling and dashboard integration come along for the ride, making it easier to track resource usage and costs across all managed clusters in a single pane of glass.
ARM64 Support Expands Platform Options
With ARM64 processors becoming increasingly common in cloud environments (thanks largely to AWS Graviton instances offering better price-performance), improved ARM64 support in v1.2.0 opens up more deployment options. The release notes mention some limitations that teams should review, but basic functionality is there for organizations wanting to take advantage of ARM-based compute.
Community Feedback Highlights Simplified Management
On a Reddit thread about large-scale Kubernetes management, user liltaf shared their experience with k0rdent: “You should look into k0rdent, it addresses all of your needs. It leverages capi but makes it very easy to use. No more complex terraform scripts to manage, it uses templates that are plain yaml files that you can GitOps. It also allows you to manage services running on each cluster and their versions using the same principle. And it is open source.”
That GitOps-friendly template approach seems to be resonating with platform teams tired of maintaining infrastructure-as-code across multiple tools and environments.
Competitive Landscape for Super Control Planes
K0rdent enters a growing market of unified Kubernetes management platforms. Cloudfleet offers CFKE with similar centralized lifecycle automation and hybrid cloud support. Rafay takes a SaaS-first approach with built-in security and lifecycle automation for public cloud, on-premises, and hybrid deployments.
What differentiates k0rdent is the open-source foundation and emphasis on avoiding vendor lock-in. The platform has been tested across AWS EC2, AWS EKS, Azure Compute, Azure AKS, vSphere, and OpenStack, with extensibility for custom providers. Mirantis also offers commercial versions—k0rdent Enterprise with 24/7 support, and k0rdent AI for machine learning lifecycle management—but the core platform remains freely available.
The project is building community momentum too. Maintainer Dina Belova led the first community call in July where users shared their deployment stories and use cases.
What Makes This Release Matter
Here’s what’s interesting about k0rdent’s approach: they’re not trying to replace Kubernetes or abstract it away entirely. Instead, they’re solving the operational problem of managing many Kubernetes clusters without forcing teams into a single vendor’s ecosystem.
For platform engineering teams supporting multiple business units or products, each with their own infrastructure preferences, that flexibility matters. You can let development teams choose between EKS, AKS, and self-managed clusters while still maintaining centralized governance and cost visibility.
The OpenStack support, despite requiring manual network setup, is particularly significant for organizations with existing OpenStack investments who don’t want to abandon that infrastructure to adopt modern container orchestration.
Whether k0rdent can compete with well-funded competitors like Rafay or established players remains to be seen, but the open-source approach and active community engagement suggest they’re building something developers actually want to use rather than something enterprises are forced to buy.
Organizations interested in trying k0rdent can start with the QuickStart guide on GitHub, where full documentation and source code are available. The complete v1.2.0 changelog details all bug fixes and feature additions.