Capabilities
Capabilities¶
The Managed Kubernetes Service (MKS) provides a robust set of capabilities for provisioning, operating, and scaling Kubernetes clusters across public cloud and on-premises environments. It simplifies the end-to-end lifecycle management of clusters with features designed for reliability, automation, and operational efficiency.
Provisioning¶
- One-click provisioning of managed Kubernetes clusters.
- Supports Highly Available (HA) and non-HA control plane configurations.
- Master and worker nodes can be provisioned on VM, Bare Metal, or other physical infrastructure.
- Option to enable SSH for master/worker nodes.
- Blueprint applied automatically during provisioning.
- Control plane nodes can be placed across regions or availability zones based on selected infrastructure SKUs.
Networking¶
- Supports Calico CNI for cluster networking.
- Supports proxy configuration for outbound network access.
- Automatic IP allocation for all node pools.
Access Control¶
- Kubeconfig is automatically generated during provisioning.
- Role-based access controls can be applied via blueprints or post-provisioning.
GPU Support¶
- GPU injection is handled through blueprints.
- Only GPU types supported in the selected blueprint can be used.
Blueprints¶
- Support for default and custom blueprint versions.
- Automatic application of blueprint-defined system components (e.g., logging, monitoring).
Scaling¶
- Worker node scaling is supported.
- Nodes are removed using LIFO (Last-In-First-Out) policy when scaling down.
- Scale operations can be triggered via the UI or APIs.
Upgrades¶
- Kubernetes version upgrades supported for worker nodes.
- Control plane upgrades are configurable during provisioning.
Automation¶
- Workflow handlers can be attached for advanced automation use cases.
- Supports provisioning through Rafay APIs, Terraform, or cloud-init.
Multi-Infrastructure Support¶
- Supports infrastructure SKUs on VM, Bare Metal, and physical nodes.
Tagging and Annotations¶
- Supports tagging for all resources (master, worker, and control plane).
- Node-level annotations supported.
Day-2 Operations¶
- View cluster and node metadata.
- Download kubeconfig for cluster access.
- Perform scale and upgrade operations.
- Attach new worker pools post-deployment.
Outputs¶
- Kubeconfig
- Master and worker node details
- Network configuration
- Private key for SSH access (if enabled)
Monitoring and Validation¶
- Region compatibility checks during provisioning.
- Validations for runtime dependencies and supported configurations.
- Access to node discovery and provisioning logs.