{
  "id": 11226616,
  "title": "A live Kubernetes cluster can still have an ownership gap",
  "url": "https://urgent.news/2026/10/01/a-live-kubernetes-cluster-can-still-have-an-ownership-gap",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-10-01T15:30:00.000Z",
  "source": {
    "name": "The New Stack",
    "slug": "the-new-stack",
    "url": "https://thenewstack.io/kubernetes-operations-ownership-governance/"
  },
  "original_language": "en",
  "account": "On the initial launch day of a Kubernetes deployment, an IT department often views the establishment of the cluster, routing of networking, and deployment of application containers as significant achievements. However, Day 2 reveals a more complex reality: the cluster may function well, but the applications running on it might not. This situation raises the question of who is responsible for ensuring the smooth operation of applications after the initial Kubernetes deployment.\n\nKarthik Subramanian, a Principal Product Manager for HPE Morpheus Software at Hewlett Packard Enterprise (HPE), gives an interview with The New Stack to clarify the responsibilities involved in post-launch Kubernetes operations. These responsibilities include access management, handling of configuration drift, and managing cluster upgrades. This conversation marks the second part in a four-part series.\n\nThe division of responsibilities between platform and application teams is crucial in ensuring clear ownership and effective governance. Platform teams, which include cloud architects, Kubernetes platform administrators, security specialists, and monitoring experts, are responsible for overseeing the overall Kubernetes environment. This includes tasks such as cluster provisioning, namespace allocation, version management, storage management, and enforcement of security policies.\n\nApplication teams, comprising core developers, QA personnel, and production deployment engineers, are responsible for deploying workloads within designated namespaces, managing application code releases, and using cluster capabilities without interfering with the underlying control plane. Although these responsibilities may be assumed by a single individual, it is beneficial to assign an owner for each function to clearly define handoffs and access decisions.\n\nSubramanian outlines two levels at which platform teams can set standards: cloud-level and cluster-level standards. Cloud-level definitions involve establishing access and role-based access control (RBAC) policies, network and admission requirements, backup and recovery objectives, infrastructure-as-code standards like Terraform, and service-level indicators for the environment.\n\nCluster-level standardizations, on the other hand, pertain to localized settings tailored to specific environment constraints. This includes specifying the Kubernetes version support cadence, naming conventions, resource labeling, multi-tenancy namespace models, and cluster-specific network policies for pod-to-pod communications.\n\nHPE Morpheus Software offers a solution to automate governance and cluster upgrades, providing platform teams with a way to connect cloud resources, Kubernetes provisioning, blueprints, workflows, and role-based access controls. Teams can configure catalog access, set quotas, define approval policies, and integrate with IT service management systems.\n\nWhen it comes to upgrading Kubernetes clusters, Subramanian suggests a staged upgrade process to minimize disruption. This process involves pre-validation checks to review node health and compatibility warnings, sequential updates of control-plane and worker nodes, and the use of canary and blue-green approaches for production changes. Traffic can be gradually shifted to newer clusters with minimal application impact, and the upgrade is considered complete only when the application's critical path functions correctly and service objectives remain within tolerance.",
  "summary": "Day 1 of a Kubernetes deployment usually feels like a milestone to an IT department: Clusters are provisioned, networking is The post A live Kubernetes cluster can still have an ownership gap appeared first on The New Stack .",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}