Monitoring network interaction between all cluster nodes, as well as (optionally) up to additional external nodes

Description

The network communication monitoring module provides continuous connectivity verification between all the main and, if necessary, external nodes of the cluster.

Module features:

• automatically checks the availability of all cluster nodes (and, optionally, external systems) using ICMP (ping) — testing is started every two seconds;
• all results are exported in metrics format to the Prometheus monitoring system.;
• included is a ready—made dashboard for Grafana, where current availability, delay schedules, and potential network connectivity issues are visualized in real time;
• allows you to quickly identify nodes with degraded connectivity and speeds up the response to incidents.

How does it work?

The module tracks the node’s .status.addresses field for changes. Upon detecting changes, it invokes a hook that collects a complete list of node names/addresses and passes it to a DaemonSet (the latter recreates the Pods). As a result, ping checks the always up-to-date list of nodes.