ScaleOps Multi-Cluster Architecture

Overview

ScaleOps provides enterprise-grade multi-cluster management capabilities, enabling centralized control and monitoring of multiple Kubernetes clusters from a single dashboard. This architecture delivers aggregated analytics, unified reporting, and streamlined cluster management across your entire infrastructure.

Deployment Architecture

Parent Cluster Configuration

Deploy ScaleOps on your designated parent cluster by following the standard Installation Guide.

Network Connectivity Requirements

Ensure that your parent cluster is accessible from all child clusters. This typically requires configuring an Ingress or load balancer to expose the parent cluster’s API.

Child Cluster Deployment

Deploy ScaleOps on each child cluster using the standard installation process with the following additional parameter:

--set parentURL=<ENTER_HERE_PARENT_URL>

Your multi-cluster environment is now configured and ready for centralized management.

Multi-Cluster Capabilities

Cluster Navigation

Once your clusters are connected, access the parent cluster’s dashboard to utilize the cluster selection dropdown. This interface enables seamless navigation between different child clusters, streamlining your cluster management workflow.

Unified Overview Dashboard

The multi-cluster overview page provides comprehensive insights across all connected clusters. This dashboard displays aggregated metrics, feature adoption progress, cost optimization opportunities, and performance insights with direct navigation to detailed analysis views.

Aggregated Analytics

Access comprehensive analytics across all clusters through the analytics interface. This unified view provides aggregated data and reports to support informed decision-making regarding cluster resources and workload optimization strategies.

Network Architecture

The multi-cluster architecture employs a hub-and-spoke model with secure WebSocket communication:

• Parent Cluster: Listens for incoming connections on port 443
• Child Clusters: Establish and maintain persistent WebSocket connections to the parent cluster
• Request Routing: When the parent receives requests targeting a child cluster, it forwards them through the corresponding WebSocket connection
• Data Locality: All metric data remains within child clusters; the parent cluster does not store or process child cluster metrics

The WebSocket connections are initiated and maintained by child clusters, requiring only the parent cluster to accept inbound connections. Once established, these connections provide bidirectional communication for request forwarding and response handling.