Hello ,
We have maximize the Logstash performance in terms of processed events per sec, by tuning vcpus, Heap, pipeline workers, Consumer Threads, N/w bandwidth, Batch size etc. There is no scarcity of resources assigned to Logstash application which is running on a VM machine. The Logstash
is reading from Kafka Topic and writing to Elastic Cluster Hot node via Coordination Nodes. We also tuned Topic partitions.
I wanted to know whether any other point is available which can improve the Logstash
performance, e.g. CPU Pining, socket sharding, Kernel bypass etc. Are these features supported by Logstash and will improve it’s performance. Any feedback on this will be quite helpful.