I’ve been wondering what is wrong with my v3.3.5 setup.
server.conf contains nodes of my Elastic cluster:
elasticsearch_hosts = http://<es-node-1-IP>:9200,http://<es-node-2-IP>:9200,http://<es-node-3-IP>:9200.
One of the Elastic nodes died yesterday. The rest of the cluster should survive. But still requests on Graylog REST API (particularly to
/api/search/universal/absolute) timed out.
Naïvely, I’d assume Graylog should round-robin to the rest of Elastic nodes. I can’t prove two remaining nodes from the cluster worked OK, but still.
Am I missing something obvious?