We are using logstash tcp/udp json input plugin and the elasticsearch output plugin. I was wondering if there was a buffer of sorts between the two, so that, if the elasticsearch instance is restarted, messages are not lost.
I have read about using redis as a broker and so forth, redis seems to be installed on our server, but I can't find any association with it and logstash. I have also come across a queue depth of 20, hardcoded it seems, and was wondering if that was some kind of buffer. If so, is it persistent so that it can survive a logstash restart? and can it be increased?