Hi,
I am using ELK stack for log monitoring. Its in production now. The pipeline I use is as follow.
FB-->Logstash-->Redis-->Logstash-->Nginx-->ES
I am now reading about kafka. Just for quick direction, please suggest how can I use kafka for better performance, better feature in my current pipeline.
Basically I want to get rid of some of the components in my pipeline. those are too much to maintain.
Filebeat can write directly to Kafka and Redis, so it should be possible to simplify the architecture and remove the first Logstash instance even if you do not switch to Kafka. For further advice, have a look at this blog post. Depending on your scale and requirements, it is worth noting that Logstash now has support for buffering using persistent queues, which can removing the need for a separate message queue altogether.
I would recommend upgrading your ingest pipeline. As far as I can see, Filebeat does now support SSL for the Redis output. Persistent queue support is available in the latest Logstash release, so I would recommend upgrading this too.
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.