I've setup an elasticsearch cluster with 3 Elasticsearch nodes, 1 Kibana, 1 Logstash. Everything is working as expected with one exception.
I currently have my Beats (metric, file, etc.) sending log data to my Logstash node. My question is, if I have to take the Logstash nodes down for maintenance, upgrade, unexpected crash, etc., what happens to the log data that the beats are collecting?
I tried to simulate this by bringing logstash offline for 30 minutes, then bringing it back online. I ended up with a 30 minute gap in metricbeat data but actually got all the filebeat data. Is this expected behaviour? Is there a way to modify this and have metricbeat behave like filebeat in this respect?