How to run Kubernetes E2E performance/density tests and obtain latency and throughput measurements?

I wanted to replicate the exact measurements and tests realized here or here (API responsiveness, pod startup time, networking/scheduling throughput ...etc.) and obtain the results to make visual graphs out of them but there is no clear guidance how to do it.

I compiled kubetest, kubectl, ginkgo and e2e.test and run the e2e tests but I don't find any way to do what has been done in the previous articles. Running with -ginkgo.dryRun shows all the list of possible tests and apprently there is no tests that match the ones I want to do (using -ginkgo.focus="Performance|Density" or other combinations doesn't also yield any interesting results).

I just want to know how the previous articles realized their tests and obtained the measurements of latency, throughput, pod startup times as well as other metrics. What benchmark they used and how they applied it ?

PS: I use a Kubernetes v1.17.0 cluster created with Kubespray on a bare-metal infrastructure (high-end servers). Used OS is Debian 10.