Chapter 7. Monitor hardware health with node-proxy agent
Red Hat Ceph Storage uses the node-proxy agent, a lightweight service running on each node, to monitor hardware health. The agent collects hardware status and inventory from Out-of-Band (OOB) interfaces, such as iLO, using the Redfish API. The agent processes the collected data locally and sends it to the Ceph Manager (MGR) daemon, which displays the results in the Dashboard. illustrates the monitoring flows.
Figure 7.1. Hardware monitoring with node-proxy agent
The node-proxy agent collects hardware data from the Redfish API and then performs the following: - Pushes the data to the Ceph Manager - Serves Ceph Dashboard queries. - Displays the hardware health information in the Ceph Dashboard.
Network connectivity and port configuration for node-proxy agent monitoring
Hardware monitoring with the node-proxy agent relies on specific communication paths and port configurations. The following connections must be established and accessible: - Ceph Manager (MGR) communicates with node-proxy agents—typically over port 9283. - Node-proxy agents connect to Out-of-Band (OOB) interfaces using the Redfish API, usually over port 443 (HTTPS).
Ensure that all required ports are open and network connectivity exists between the MGR, node-proxy agents, and OOB interfaces to enable reliable data collection and reporting.
Limitations
This feature has the following limitations:
- The node-proxy agent depends on the Redfish API, which requires reliable access to the OOB management network.
- Both the node-proxy agent and the ceph-mgr daemon must have network connectivity to the OOB interfaces for monitoring to function correctly.