Benchmarking is hard, especially on a macro level that integrates multiple code components into one or multiple microservices. It’s challenging to reproduce production conditions that matter, isolate the test from external effects and reliably automate all steps. This leads to unreliable benchmark results, a long time to execute benchmarks and a high cost of stress tests, so generally wasted engineering time and business money.
Fortunately, with good tools and some education, we can turn macro benchmarks into remarkably effective tools that catch performance regressions and let us know where our workloads should be optimized further!
In this talk, Bartek, an author of the “Efficient Go” book with O’Reilly and Prometheus maintainer, will walk you through the basics of macro benchmarks and load tests. You will learn about open-source frameworks, observability tools and patterns you can use to benchmark your applications in Cloud, on Kubernetes and beyond. Gain significant efficiency improvements by Macro Benchmarking correctly and with ease!