Overview

Performing a benchmark using kube-burner is relatively simple. However, it is sometimes necessary to analyze and be able to react to some KPIs in order to validate a benchmark. That is why kube-burner ships metric-collection and alerting systems based on Prometheus expressions.

Kube-burner also ships an indexing feature that, in combination with the metric-collection and alerting features, can be used to analyze these KPIs in an external tool, such as Grafana or similar.

The benchmark stages include the following:

flowchart TD
    A[/"Init benchmark"/] -- Read config --> B(Start measurements)
    B --> C(Run Job)
    C --> D{Next job?}
    D --> |Yes| C
    D --> |No| E[Stop measurements]
    E --> F[Evaluate alerts]
    F --> G[(Index results)]
    G --> H[Indexing]
    H --> I*[/End/]