**Rancher** is an open source multi-cluster [[Kubernetes]] management platform from SUSE. It provides a single control plane to deploy, manage, and monitor multiple Kubernetes clusters — whether they run on bare metal, on-premises VMs, or across cloud providers. --- ### First Principle: Managing one Kubernetes cluster is hard. Managing twenty is an entirely different problem. At hyperscaler scale, you're running dozens or hundreds of clusters across multiple regions. Rancher is the layer that makes that manageable: one UI, one RBAC model, one fleet deployment tool, across all clusters. --- ### Key Considerations - **Cluster Provisioning**: Rancher can provision new clusters on bare metal (via RKE2/K3s), on VMs ([[OpenStack]], vSphere), or adopt existing clusters. The Rancher Management Server sits outside the workload clusters. - **Fleet (GitOps)**: Rancher's Fleet tool deploys applications across many clusters from Git repositories — scaling [[ArgoCD]]-style GitOps to hundreds of clusters simultaneously. - **Multi-Tenancy**: Rancher Projects provide a namespace-based multi-tenancy model within clusters. Combined with downstream cluster isolation, it provides reasonable tenant separation for a private cloud. - **RKE2 / K3s**: Rancher maintains RKE2 (hardened, FIPS-compliant Kubernetes distribution) and K3s (lightweight Kubernetes for edge/IoT). - **Observability Integration**: Rancher integrates [[Prometheus]], [[Grafana]], and Alertmanager per cluster via the Rancher Monitoring app. - **vs [[OKD]]**: OKD is more opinionated and enterprise-focused. Rancher is more flexible and cloud-agnostic. --- ### How It Fits ``` Rancher Management Server (single pane of glass) → Multiple [[Kubernetes]] clusters (RKE2, K3s...) → Fleet (GitOps deployment) → [[Prometheus]] / [[Grafana]] (per-cluster monitoring) ``` [[Kubernetes]] | [[ArgoCD]] | [[Prometheus]] | [[Grafana]] | [[OKD]] | [[Open Source Hyperscaler MoC]]