IBM Storage Ceph Multi-Cluster Manager overview

Edit online

IBM Storage Ceph Multi-Cluster Manager provides centralized monitoring, management, and upgrade operations for multiple IBM Storage Ceph clusters.

The IBM Storage Ceph Multi-Cluster Manager dashboard provides a centralized view of cluster health, capacity, performance, alerts, software versions, and cluster management tasks.

The Multi-Cluster Manager consists of a server component and an agent that runs on each managed cluster. The server hosts the dashboard, APIs, monitoring services, and management functions, while the agent communicates with the local Ceph cluster and reports information to the server.

Figure 1 illustrates the Multi-Cluster Manager architecture and data flow.

Figure 1. Multi-Cluster Manager architecture flow

IBM Storage Ceph Multi-Cluster Manager server (application)

The IBM Storage Ceph Multi-Cluster Manager server provides the core services for centralized monitoring and management.

Install the Multi-Cluster Manager server on a dedicated management node or virtual machine that can be connected to by all participating IBM Storage Ceph clusters. The Multi-Cluster Manager server hosts the following services:

Multi-Cluster Manager API
Provides an interface for processing user requests and system operations.
Nginx reverse proxy
Handles TLS termination and request routing to internal services. The Nginx reverse proxy uses TLS certificates for secure communication. If the default certificate expires or must be replaced with a CA-signed or renewed certificate, perform the replacement on the Multi-Cluster Manager server node by using the certificate replacement script: 
/scripts/replace-certificate.sh --cert <cert path> --key KEY_PATH
Prometheus and AlertManager
Metrics received from the cluster Prometheus instances through remote_write.
Multi-Cluster Manager dashboard
Provides a unified view of cluster health, capacity, performance, alerts, software versions, and cluster management operations.
Important: For production environments, deploy the IBM Storage Ceph Multi-Cluster Manager server as a dedicated Red Hat Enterprise Linux (RHEL) Image Mode instance to keep it independent of Ceph cluster upgrades.

IBM Storage Ceph Multi-Cluster Manager agent

The IBM Storage Ceph Multi-Cluster Manager agent runs on each managed IBM Storage Ceph cluster and is deployed automatically when a cluster is attached to Multi-Cluster Manager.

The agent gathers cluster and node information from the local Ceph environment and securely transmits it to the Multi-Cluster Manager server. Performance and capacity metrics from each cluster are stored in the centralized Prometheus instance on the IBM Storage Ceph Multi-Cluster Manager server, where alerts are evaluated and categorized by severity.

The agent provides the following capabilities:

  • Collects cluster topology, health, and configuration data from the local Ceph dashboard API.
  • Pushes the collected data securely to the Multi-Cluster Manager API using TLS.
  • Configures Prometheus remote_write on the cluster to stream metrics to the central Prometheus instance.
  • Sends periodic heartbeat updates to report cluster and agent health status.
Note: All communication between the Multi-Cluster Manager agent and Multi-Cluster Manager API is secured through TLS.