How LivePerson cut observability pipeline costs by benchmarking Google Cloud infrastructure

This post details LivePerson's methodology for optimizing their observability pipeline costs on Google Cloud by benchmarking Logstash and Kafka performance across different machine types and compression codecs. Key findings include the superior cost-efficiency of n4d-standard-2 instances for Logstash and the performance gains from using LZ4 compression for Kafka. The post emphasizes the importance of infrastructure selection as a high-leverage cost optimization decision at scale and recommends periodic re-evaluation of cloud infrastructure choices.