The most elegant way to performance test your microservices running on Kubernetes

“If you cannot measure it, you cannot improve it.” — Lord Kelvin

Application programming interfaces ( APIs) are the core system of most services. Client, web, and mobile applications are all built from APIs. They sit on the critical path between an end-user and a service, and they’re also used for intra-service communication.

Because APIs are so critical, API performance is also essential. It doesn’t matter how well-built your front-end application is if the API data sources it accesses take several seconds to respond. This is especially true in a world of microservices, where services depend on each other to provide data. In my opinion, the best feature your API can offer is great performance.

To measure API performance, you need to benchmark your APIs as reliably as possible, which can be challenging. The optimal approach depends on your performance objectives. In this article, I’ll guide you through an elegant process for measuring the performance of backend applications running on Red Hat OpenShift or Kubernetes. You’ll also learn how to use Vegeta, a versatile HTTP load testing and benchmarking tool written in Golang. We will deploy Vegeta on OpenShift and run performance tests in both standalone and distributed modes.