Skip to content

Benchmarks

Supported from version zooz/predator:1.3.0

Motivation#

Load test results increase in their value and importance when compared to a certain benchmark.

By creating a benchmark for a specific test, each subsequent test run for that test will be given a score from 0-100 summarizing the test run in one simple to analyze numerical value.

The benchmarks feature alongside the compare test runs are tools that leverage predators reports and ensures that new releases stand up to your performance requirements.

Setting Up#

Test as a Benchmark#

To set a test as a benchmark, choose a test run that you are satisfied with the results, open its report and click on Set as Benchmark. Now, all future runs of this test will have their score calculated based on the benchmarks selected.

Benchmark Weights#

There are 5 factors taken from the test run results that affect the score of the run.

  • Median: Percentage of the score affected by median results
  • P95: Percentage of the score affected by p95 results
  • Server errors: Percentage of the score affected by server errors ratio
  • Client errors: Percentage of the score affected by client errors ratio
  • RPS: Percentage of the score affected by requests per second results

Please refer to the configuration manual for further documentation.