In Logstash under the pipeline directory there will be multiple .conf files that set specific grok filters. The format of the files are as follows:
Input
Grok Filter
Output [list of nodes]
For that Output section do you need to put every node that is listed in your Cluster? Or instead, would you only list the dedicated Ingest nodes as the Output?
Awesome, thank you! If those dedicated ingest nodes are listed on the Output on logstash. Would I then need to make any additional changes for the data to go from the ingest nodes to the data nodes? Or would elastic take care of that on its own?
Thanks, that's a huge help. I really appreciate it. Only quick follow up. In the environment, when we do the API command "GET /_ingest/pipeline" we see grok filters are running as pipeline processors on the Ingest nodes. If we have the complete ELK stack would a more default/generic solution be to just let these run on Logstash instead? Ive been researching and it shows that usually its either Logstash or Ingest and not very often do people have both of them combined. Is that correct?
OK cool, options are always a good thing. This was a super fast and helpful reply. Definitely going to start another thread about something else lol. Thanks again.
Sorry, a question popped up at work. What would happen if you list a node in that Logstash output section, but it is NOT an ingest node. Any loss of data? Would some data not get run though the pipeline processors? Etc (we have deployments that come with defaults so just want to cover our basis if anything could go wrong if that output node listing is wrong).
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.