A tick message is a periodic liveness probe exchanged between connected nodes. Each node uses these probes to check whether its peers still respond. section 12.1.1
Tick messages establish practical failure behavior in a distributed cluster:
- A peer that fails to respond to four consecutive ticks is considered disconnected and removed from the connected-node list. section 12.1.1
- A lost connection can indicate a network partition, including a split-brain situation where separate clusters continue serving clients. section 12.2.5
- Tick failure does not automatically reconnect the node; applications can observe connection changes with Node.monitor/2 or :net_kernel.monitor_nodes/1. section 12.1.1 section 12.2.5
$ iex --sname node1@localhost iex(node1@localhost)1> :net_kernel.monitor_nodes(true)
Subscribe a process to node-connection notifications, then observe nodeup and nodedown messages as peers connect or disappear. section 12.2.5
