Watch the Recordings: Mirantis Talks about Platform Engineering and GitOps at KubeCon NA 2025

KubeCon + CloudNativeCon North America 2025 showcased the latest innovations and best practices for platform engineering and GitOps, with new patterns and tooling shaping how Kubernetes platforms are built and operated. Mirantis speakers took to the stage in Atlanta to highlight practical strategies for reducing complexity across sprawling environments while enabling teams to move faster with confidence. From simplifying cluster lifecycle management with Cluster API to modernizing workflows through multi-repository architecture, these sessions provided guidance to organizations on how to standardize infrastructure and deliver applications more reliably at scale. Here’s a closer look at the ideas and insights our speakers shared about platform engineering and GitOps.  

Introduction to Cluster API

Mirantis Senior Principal Engineer Jussi Nummelin explains how Cluster API provides a declarative, Kubernetes-native framework for provisioning, upgrading, and operating Kubernetes clusters at scale. Cluster API uses a centralized management cluster to control fleets of workload clusters across heterogeneous infrastructure. By defining clusters through extensible custom resources and pluggable providers, teams gain a consistent model for creating and standardizing clusters while abstracting away provider-specific complexity. This architecture mirrors core Kubernetes patterns and enables scalable operations, automated remediation, and GitOps-driven workload management. This approach helps organizations overcome cluster sprawl, enforce consistent configuration, and build repeatable platform engineering workflows for multi-cluster environments. 

Key Takeaways:

• Cluster API Provides a Unified Model for Multi-Cluster Operations: A management cluster becomes the central place to provision, upgrade, and operate Kubernetes fleets through declarative APIs and custom resources

• A Pluggable Provider Framework Enables Consistent Management Across Any Environment: Infrastructure, control plane, and bootstrap providers let teams run clusters on clouds, bare metal, or Kubernetes-native control planes using the same workflow

• Workload Management Stays Flexible and Scalable: Teams can use GitOps for individual clusters or centrally manage add-ons with tools like k0rdent, a composable Kubernetes management platform, to reduce configuration sprawl

Using Multiple Repos for an Argo CD Application? It’s Not as Complicated as You Think 

Mirantis Developer Advocate Anjelica Ambrosio outlines how multiple repositories create a more modular, secure, and scalable GitOps model for teams using Argo CD. As organizations adopt multiple clusters and environments, a mono-repo can lead to slow pipelines, merge conflicts, operational risk, and security exposure. A multi-repo, on the other hand, improves clarity of ownership and reduces cross-team interference. This architecture strengthens access control, simplifies audits, isolates sensitive information, and improves versioning and rollbacks. When paired with k0rdent, a multi-repo Argo CD setup enables end-to-end automation, scalable cluster and application delivery, and a GitOps-native workflow across diverse environments.  

Key Takeaways:

• Mono-Repos Become Bottlenecks: Shared pipelines, global test runs, merge conflicts, and security exposure make a mono-repo increasingly risky and inefficient as teams grow

• Signals for Migration: Slow pipelines, risky rollbacks, poor auditability, and clashes in code changes are all signs that a mono-repo is no longer fitting a team’s needs

• Multi-Repo Architecture Provides Security: Granular permissions, isolated histories, and repository-level boundaries reduce risk and simplify audits; meanwhile, separate repositories create modular boundaries that support safer changes and reduced blast radius during incidents

See You at the Next Event!

Throughout the platform engineering and GitOps sessions at KubeCon Atlanta, Mirantis speakers emphasized that success depends on adopting consistent, automated, and secure patterns for managing both clusters and applications. Declarative lifecycle management with Cluster API, combined with clean separation of concerns through multi-repo GitOps workflows, gives teams the structure needed to operate at scale without sacrificing velocity. 

If you didn’t get a chance to meet us in Atlanta, we’ll also be at KubeCon + CloudNativeCon EU in Amsterdam in March 2026. Hope to see you there!