We should have a basic benchmark framework for tracking performance variation across versions. Additional features should be added as follow-ups.
- The benchmark framework should include:
- A set of name servers serving a set of zones
- A CLI tool that runs Zonemaster Engine on a set of zones and produces a report
- A list of zones to run the CLI tool on
- An instruction for running the benchmark.
- The report should be in a machine-readable format and should include:
- Benchmark start time
- Machine hostname
- Zonemaster Engine version
- Average per-zone duration
- Individual per-zone durations
Perhaps the test-zone-data could be reused for this purpose. While that data is certainly not representative of the real-world, we do control it, and it is readily available.
N.B., MVP is an abbreviation of Minimum Viable Product.
We should have a basic benchmark framework for tracking performance variation across versions. Additional features should be added as follow-ups.
Perhaps the test-zone-data could be reused for this purpose. While that data is certainly not representative of the real-world, we do control it, and it is readily available.
N.B., MVP is an abbreviation of Minimum Viable Product.