Closed r0h4n closed 7 years ago
@brainfunked Need your inputs
@anmolbabu mentioned tendrl-performance-monitor ping's the tendrl managed nodes frequently, we need to figure out if the "tendrl/monitor" node-agent should monitor all nodes or tendrl-perfromance-monitoring should do that.
tendrl-node-agent provide's heartbeats to central store, however this requires external monitoring to ensure all node-agents are running.
Currently, tendrl node-agents running on the tendrl-api node are tagged with "tendrl/monitor" role.
Members of this role will monitor all tendrl-node-agents. In case of a dead or zombie node-agent, the monitoring node-agents will mark the same appropriately (i.e. "unavailable")