This article is based on reporting and analysis originally published on Forbes.com around Microsoft Build 2020, when Azure Arc-enabled Kubernetes entered public preview.
What Is Azure Arc?
Azure Arc is Microsoft’s hybrid and multi-cloud management platform. First announced at Microsoft Ignite 2019, it extends the Azure control plane to resources running outside of Azure — including on-premises data centers, edge locations, and competing public clouds such as AWS and Google Cloud.
At Build 2020, Microsoft moved the first two pillars of Azure Arc into public preview: management of servers (VMs) and management of Kubernetes clusters running outside of Azure. A third pillar — Azure Arc-enabled data services — remained in private preview at that time.
At a high level, Azure Arc delivers three capabilities:
- Arc-enabled Servers — Register and manage Linux and Windows machines hosted anywhere as Azure resources.
- Arc-enabled Kubernetes — Attach Kubernetes clusters running on-premises or in other clouds to the Azure control plane.
- Arc-enabled Data Services — Run managed Azure SQL and PostgreSQL Hyperscale instances inside any Arc-registered Kubernetes cluster.
Azure Arc for Servers: Bring Your Own Machine to the Cloud
Azure Arc for Servers allows customers to connect physical or virtual Linux and Windows machines hosted outside of Azure directly to the Azure control plane. Once connected, each machine is treated as a native Azure resource — it receives an Azure resource identifier, belongs to a resource group under a subscription, and can be managed with standard Azure tooling.
How Onboarding Works
Onboarding a server requires installing the Azure Connected Machine agent, available for both Windows and Linux. Once installed, the agent sends a periodic heartbeat to the Azure control plane to signal availability and readiness to receive configuration changes.
At the time of the Build 2020 announcement, supported operating systems included:
- Windows Server 2012 R2 and later
- Ubuntu 16.04
- SUSE Linux Enterprise Server 15
- Red Hat Enterprise Linux 7
- Amazon Linux 2
Support for SUSE Linux Enterprise Server and Red Hat Enterprise Linux was announced at Build 2020. The Connected Machine agent is updated via Microsoft Update on Windows; on Linux, upgrades are performed manually.
Managing Hybrid Machines as Azure Resources
After registration, Arc-connected machines appear in the same Azure resource group alongside native Azure VMs. Teams can apply Azure tags, enforce Azure Policy, and audit compliance across all machines in the group — regardless of whether those machines are running in Azure, a private data center, or another cloud provider.
A practical example: an enterprise IT team can define an Azure Policy that ensures all Linux machines in a resource group only allow SSH key-based authentication and automatically disables password-based accounts. The policy applies uniformly to every registered machine in the group, regardless of physical location.
Azure Arc-Enabled Kubernetes: Bring Your Own Cluster to the Cloud
Azure Arc-enabled Kubernetes follows the same philosophy as Arc for Servers, but targets Kubernetes clusters instead of individual machines. Clusters running anywhere — bare-metal on-premises, managed services like Amazon EKS or Google Kubernetes Engine, or enterprise distributions like Red Hat OpenShift and VMware Tanzu Kubernetes Grid — can be attached to Azure Arc.
The Arc Agent in Kubernetes
When a cluster is registered with Azure Arc, an agent is deployed under the azure-arc namespace. Similar to the Connected Machine agent for servers, this agent watches for configuration requests from the Azure control plane. It also sends telemetry to Azure Monitor, including:
- Cluster inventory
- Kubernetes events
- Container stdout and stderr logs
- Node, container, Kubelet, and GPU performance metrics
GitOps-Based Configuration Management
One of Azure Arc-enabled Kubernetes’s most significant features is its integration with GitOps-based configuration management. GitOps treats configuration as code: cluster state is declared in a Git repository (such as GitHub or GitLab), and a deployment agent inside the cluster continuously reconciles actual state with desired state.
Microsoft chose Flux, an open-source GitOps tool, as the GitOps engine for Azure Arc. At the time of the announcement, Flux was a Cloud Native Computing Foundation (CNCF) sandbox project. Using Flux, cluster administrators can centrally manage tasks such as:
- Namespace creation
- Role-Based Access Control (RBAC) configuration
- Container registry secrets
- Application rollouts to hundreds of clusters with a single Git commit
Azure Arc monitors configuration drift and automatically applies corrective changes to bring clusters back to their declared desired state. For example, a cluster administrator can enforce a policy preventing privileged pods from running in specific namespaces, applied uniformly across all registered clusters.
Three Core Capabilities of Arc-Enabled Kubernetes
- Global inventory management — Onboard and view all Kubernetes clusters from any deployment location in a single pane of glass.
- Centralized workload management — Deploy applications and configuration changes to hundreds of clusters simultaneously via a single Git commit.
- Policy-driven cluster management — Govern and audit cluster compliance centrally using Azure Policy.
At Build 2020, Microsoft announced partnerships with Red Hat, SUSE, and Rancher to bring Red Hat OpenShift, SUSE CaaS Platform, and Rancher Kubernetes Engine as officially supported distributions on Azure Arc.
Azure Arc-Enabled Data Services
The third pillar of Azure Arc — data services — was still in private preview at the time of the Build 2020 announcement. It aims to allow customers to run managed Azure SQL and PostgreSQL Hyperscale instances inside any Kubernetes cluster registered with Azure Arc. This would enable fully managed database capabilities anywhere the Arc agent can be deployed, including air-gapped or edge environments.
How Azure Arc Fits Into Microsoft’s Broader Hybrid Strategy
Azure Arc occupies a specific position within Microsoft’s end-to-end hybrid portfolio. Alongside Azure IoT Edge (for edge device management) and Azure Stack Edge (for on-premises Azure hardware), Arc addresses the data center and multi-cloud management layer. Together, these three products cover the full spectrum from far edge to data center to public cloud.
Microsoft’s approach with Arc differs from traditional hybrid strategies that rely on dedicated hardware or proprietary hypervisors. Instead, Arc extends the Azure control plane outward using lightweight agents, making any machine or cluster a manageable Azure resource regardless of where it runs. This reduces the barrier to adoption significantly — onboarding a VM or a Kubernetes cluster can be accomplished with a single command.
Key Takeaways for DevOps and Platform Engineering Teams
- Azure Arc enables a single Azure control plane across on-premises, edge, AWS, GCP, and Azure workloads without requiring migration.
- The agent-based model keeps onboarding lightweight — no proprietary hardware or hypervisor lock-in is required.
- GitOps via Flux provides a declarative, auditable approach to managing configuration and policy across large fleets of Kubernetes clusters.
- Azure Policy can be applied uniformly across hybrid and multi-cloud resources, simplifying compliance and governance at scale.
- Arc-enabled data services represent a significant potential differentiator, enabling managed database capabilities in any Kubernetes environment — including those running on-premises or in competing clouds.
For enterprises already invested in the Azure ecosystem, Azure Arc offers a pragmatic path to extending cloud-native management practices to infrastructure they cannot or do not want to migrate to Azure.