The post compares the performance of Rust (using the Actix framework) and Go (using the Fiber framework) in a production-ready EKS cluster on AWS. It evaluates their CPU usage, memory usage, client latency, and availability under different loads. The benchmarks include a straightforward JSON payload test and a real-world scenario involving file reads, uploads to S3, and database writes in PostgreSQL. Rust generally showed higher CPU usage and latency compared to Go. The post also discusses the importance of using Prometheus metrics such as Histograms over Summaries for scalability and accurate latency tracking.
•20m watch time
1 Comment
Sort: